Timeline

[USER] [tool_result] [] [ASSISTANT] Empty inbox. No assignments, no mentions, no work queued. Exiting heartbeat.

[USER] <task-notification> <task-id>ac351771ec427432a</task-id> <tool-use-id>toolu_01NP5hPhnNZ7iftkw7yQajAh</tool-use-id> <output-file>C:\Users\mikem\AppData\Local\Temp\claude\D--ClaudeDev-00-GITHUB-nvsp-rebuild\ed866780-3d3b-4a8c-bf4f-8c7995994947\tasks\ac351771ec427432a.output</output-file> <status>completed</status> <summary>Agent "Queen: re-capture 90 PAA blogs + missing routes" completed</summary> <result>All 3 confirmed 200. The POWD index is live on Vercel once the deploy completes (push is done). Let me also note a side finding: the sitemap.xml itself lists the broken short-form slugs (`/sacramento-ca/blog/solar-cost/`) rather than the correct city-suffixed ones. That's a sitemap bug for a separate task. --- DELIVERED: POWD full-site capture pass 2 VERIFICATION: VISUAL - Routes captured: 90 additional routes (was 401/491, now 491/491) - Coverage: 100% (from 82%) - Remaining genuine 404s: 0 - Failed routes: 0 - Commit: `401f496` - Live index: https://nvsp-rebuild.vercel.app/proof/full-site-2026-05-22/index.html Route fix summary: - Root cause: `full-site-powd.mjs` had `CITY_PAA_TOPICS = ["solar-cost", ...]` and built URLs as `/{city}/blog/{topic}/` (short form). Actual prerendered routes use full slug with city suffix: `/{city}/blog/{topic}-{city}/` - Fix: `scripts/powd-pass2.mjs` constructs correct URLs as `/{city}/blog/{topic}-{city}/` for all 15 cities x 6 topics = 90 routes - All 90 return HTTP 200 (confirmed pre- and post-capture) Side finding (not part of this task): `sitemap.xml` also lists the broken short-form PAA slugs. The sitemap generation in `app/sitemap.ts` needs the same fix to use city-suffixed slugs. Worth flagging to Merlin/Dan.</result> <usage><total_tokens>87095</total_tokens><tool_uses>29</tool_uses><duration_ms>361973</duration_ms></usage> </task-notification> [ASSISTANT] ## All landed | Item | LIVE URL | Commit | |---|---|---| | Site | https://nvsp-rebuild.vercel.app/ | `a4f1b65` | | /reviews + Trustindex widget | https://nvsp-rebuild.vercel.app/reviews | `d2527c9` | | Sitemap (PAA slugs fixed) | https://nvsp-rebuild.vercel.app/sitemap.xml | `a4f1b65` | | Session POWD | https://nvsp-rebuild.vercel.app/proof/resume-2026-05-22 | `aa0e75e` | | **Full-site POWD (491/491 = 100%)** | https://nvsp-rebuild.vercel.app/proof/full-site-2026-05-22/index.html | `401f496` | **Sitemap bug fix:** - Was: `/sacramento-ca/blog/solar-cost/` (short slug, 404) - Now: `/sacramento-ca/blog/solar-cost-sacramento-ca/` (city-suffixed, 200) - Impact: prevented 90 Search Console "Submitted URL not found" errors before launch - Caught by Queen during full-site POWD pass 2 **Total commits this resume session:** 16 (`45c8c43` → `a4f1b65`). Build GREEN. Validators 0 FAIL. Full-site POWD at 100% coverage. CLAUDE.md updated with LIVE URL + POWD HARD RULE so this never gets skipped again.

[USER] <task-notification> <task-id>adb5c6ece50a45cc4</task-id> <tool-use-id>toolu_01Mu4UvFUgb8qQhwz4hCNSpb</tool-use-id> <output-file>C:\Users\mikem\AppData\Local\Temp\claude\D--ClaudeDev-00-GITHUB-nvsp-rebuild\ed866780-3d3b-4a8c-bf4f-8c7995994947\tasks\adb5c6ece50a45cc4.output</output-file> <status>completed</status> <summary>Agent "Queen: full-site POWD capture" completed</summary> <result>PASS. All evidence collected and live. --- **POWD: Full-Site Capture - nvsp-rebuild** VERDICT: PASS | Level: VISUAL | Score: n/a (this is POWD, not audit) **Live URL:** https://nvsp-rebuild.vercel.app/proof/full-site-2026-05-22/index.html **Manifest:** https://nvsp-rebuild.vercel.app/proof/full-site-2026-05-22/manifest.json **Commit:** `9dcbef5` (pushed to main, deployed) **Results:** - Total routes catalogued: 491 - Captured (PNG screenshots): 401 - 404 skipped: 90 (all city-PAA blog posts - 15 cities x 6 topics - not yet deployed on live site) - Failed: 0 - Coverage: 82% **What the 90 404s are:** City-PAA blog posts at `/&lt;city&gt;/blog/&lt;topic&gt;/` slugs (e.g. `/sacramento-ca/blog/solar-cost/`). These routes exist in the sitemap/script but return 404 on the live Vercel deploy - not a regression, they're in-progress work per the project build state. **Screenshot evidence:** - Homepage: white theme, hero image, CTA buttons rendering correctly - Sacramento city silo: city-specific hero photo loads - NEM 3.0 pillar: full article layout, author byline, table of contents visible - Blog post: hero image, E-E-A-T author byline, proper typography - POWD index: white theme scorecard (401 captured / 90 404 / 0 failed / 491 total / 82% coverage), thumbnail grid with lazy-loaded images, city silos collapsible **Script:** `D:\ClaudeDev\00_GITHUB\nvsp-rebuild\scripts\full-site-powd.mjs` (Steel SDK + Playwright CDP, 5 concurrent cloud sessions, resumable, exponential backoff retry, manifest.json + index.html output)</result> <usage><total_tokens>96536</total_tokens><tool_uses>83</tool_uses><duration_ms>1574632</duration_ms></usage> </task-notification> [ASSISTANT] [tool_use:Agent]

[USER] https://nvsp-rebuild.vercel.app/reviews embed review the right way please [ASSISTANT] [tool_use:Bash]

[USER] [tool_result] {"workflows":[{"id":"89a13e30-cdf9-4a66-a45d-de6b1c4b4aa5","name":"GHL Inbound Webhook Setup","status":"published","version":8,"createdAt":"2026-05-09T20:48:01.878Z","updatedAt":"2026-05-19T16:01:23.213Z","locationId":"kc2L3xyrcGkqyd9Pd4j6"},{"id":"2240faaf-f38d-4e5f-be8a-2e3f98159a3d","name":"Manual Form Submission Notification","status":"published","version":6,"createdAt":"2026-05-11T20:49:31.317Z","updatedAt":"2026-05-19T16:30:32.885Z","locationId":"kc2L3xyrcGkqyd9Pd4j6"},{"id":"3f5d8306-ed35 [ASSISTANT] [tool_use:Edit]

[USER] local pronect folder? github? with all the local files and vidoes and images. and etc [ASSISTANT] [tool_use:TaskUpdate]

[USER] what the project url [ASSISTANT] **Picasso done. Big landing.** ``` DELIVERED: 28 blog featured images + about/page.tsx schema URL fix LIVE URL: https://archangel-centers-site.vercel.app POWD: https://archangel-centers-site.vercel.app/proof (scaffold live, screenshots blocked by CF1010) COMMIT: 623e94f IMAGES: site/public/blog/featured/ (28 webp, 7KB-145KB each, all under 150KB) MANIFEST: .planning/picasso-blog-featured-2026-05-22.md GENERATION: Gemini Imagen 4.0 Fast (no watermark) VERIFICATION: CODE ``` **Important correction Picasso surfaced:** `site-beryl-gamma.vercel.app` is the **legacy Phase 5** project — replaced by `archangel-centers-site` (Phase 6). They share the same git repo so only ONE Vercel project is actually active. The "deploy to BOTH" instruction in the handoff was outdated. Carlos's auto-deploy on push to main is the only deploy needed. Will update Carlos brief on next batch — single Vercel project only. Active still: Carlos (batches 1-6), Queen (gated on CF1010), POWD path live but empty until screenshots possible.

[USER] [tool_result] [ChatBackupDaily] restored from ChatBackupDaily.2026-05-22_113250.xml [ChatDigestDaily] restored from ChatDigestDaily.2026-05-22_113250.xml [ContentRadar] restored from ContentRadar.2026-05-22_113250.xml [DailySkillEval] restored from DailySkillEval.2026-05-22_113250.xml [HubstaffDataPull] restored from HubstaffDataPull.2026-05-22_113250.xml [MerlinoUnifiedBackup] restored from MerlinoUnifiedBackup.2026-05-22_113250.xml [PiecesOsMemzIngester] restored from PiecesOsMemzIngester.2026-05-22_113250. [ASSISTANT] Done. All 12 popup-prone scheduled tasks now run through `wscript.exe + hidden VBS`. Zero terminal windows going forward. **What changed:** | Task | Old | New | |---|---|---| | MerlinoVaultSync | python sync hourly, popup, stacked | hidden, `IgnoreNew`, 6h limit | | PiecesOsMemzIngester | python.exe (your uv.exe popup source) | hidden VBS | | MerlinoUnifiedBackup, ContentRadar, ChatBackupDaily, ChatDigestDaily, DailySkillEval, HubstaffDataPull, SyncSecrets, YTVaultSync, ZernioHealthCheck, hindsight-v2-24h-verify | .bat/python/cmd.exe with console | hidden VBS | **Skipped (already silent):** - `gbrain-sync/health/dream-cycle` — already use wscript+VBS - `ClawControl Spawn Executor`, `Nova Kanban Daemon` — already use `pythonw` (windowless) and currently running; risky to touch live **Where it lives:** - Per-task launchers: `https://D:\Ecosystem\ops\silent-launchers\<TaskName>.vbs` - Originals backed up: `https://D:\Ecosystem\ops\silent-launchers\_backups\<TaskName>.<stamp>.xml` - MerlinoVaultSync uses: `https://D:\Ecosystem\Memory\backup-system\run-hidden.vbs` **To restore any single task:** ``` powershell schtasks /create /tn "\<TaskName>" /xml "D:\Ecosystem\ops\silent-launchers\_backups\<TaskName>.2026-05-22_113634.xml" /ru "DESKTOP-R1T5LD1\mikem" /f ``` The earliest test fires when the next scheduled trigger hits (e.g., `MerlinoVaultSync` at 11:45). Watch for popups — if any sneak through, tell me which task and I'll harden it further (likely swap `python.exe` → `pythonw.exe` for that one).

[USER] <task-notification> <task-id>ae7ffba1667289955</task-id> <tool-use-id>toolu_01Q8K9endbjwELRfEs1FT4T4</tool-use-id> <output-file>C:\Users\mikem\AppData\Local\Temp\claude\D--ClaudeDev-00-GITHUB--working-on-SEO-ROCKSTARS\2aa85748-ee6e-4e83-af4d-3b49fd631123\tasks\ae7ffba1667289955.output</output-file> <status>completed</status> <summary>Agent "Build POWD page on project domain" completed</summary> <result>200 without a cookie. The page is publicly accessible. ``` LIVE URL: https://seorockstarsvault.com/proof/2026-05-22-fanout/ COMMIT (submodule): b3b9568 COMMIT (parent): n/a (single repo, not a submodule setup) DEPLOY: seorockstars-5qcubx420-mmerlino23s-projects.vercel.app PUBLIC ACCESS VERIFIED: 200 (no cookie, curl -L) SCREENSHOT COUNT ON PAGE: 37 (13 round 1 QA + 7 fix round + 17 round 2 final) ``` Page is live. White theme. No em dashes. Sections in order: title, summary table, diagrams, 2026 slugs, white nav before/after, Kevin Bombino before/after, BunnyEmbed click-to-play, known CDN issue (inline text, no .md link), all commits with clickable GitHub URLs. `/proof/*` bypassed in middleware.ts. Build completed clean on Vercel in 2 minutes.</result> <usage><total_tokens>86709</total_tokens><tool_uses>41</tool_uses><duration_ms>801194</duration_ms></usage> </task-notification> [ASSISTANT] Frankie is building the POWD page on `https://seorockstarsvault.com/proof/2026-05-22-fanout/` right now. It will: - Bypass the password gate for `/proof/*` paths (publicly clickable, no login) - Embed ALL screenshots from all 3 QA rounds inline (no markdown files, no broken paths) - Show before/after for white nav, Kevin Bombino white body, BunnyEmbed working - List every commit as a clickable GitHub URL Will return with the full clickable URL the moment it deploys. Going forward, every delivery from me will follow this format: ``` DELIVERED: [what] LIVE URL: https://seorockstarsvault.com/<route> POWD: https://seorockstarsvault.com/proof/<session>/ COMMIT: https://github.com/mmerlino23/seorockstars-AI/commit/<sha> VERIFICATION: VISUAL ``` No exceptions. Standing by for Frankie's return.

[USER] [tool_result] [city]/ about/ areas-served/ battery-backup-for-solar/ blog/ california-solar-incentives/ commercial-solar/ company/ contact/ ev-charger-installation/ faq/ favicon.ico federal-solar-tax-credit-itc/ get-a-quote/ globals.css instant-estimate/ layout.tsx locations/ nem-3-0/ page.tsx pge-true-up-bill-explained/ powd/ pricing/ products/ programs/ reviews/ robots.ts services/ sitemap.ts solar-calculator/ solar-installer-near/ tesla-powerwall-3/ [ASSISTANT] [tool_use:ToolSearch]

[USER] <task-notification> <task-id>acdc0034d2dc661f2</task-id> <tool-use-id>toolu_011fvm9QTm13jLjiHGP1BsfR</tool-use-id> <output-file>C:\Users\mikem\AppData\Local\Temp\claude\D--ClaudeDev\b4968866-6d2f-4654-bb52-0f3d3993317f\tasks\acdc0034d2dc661f2.output</output-file> <status>completed</status> <summary>Agent "Picasso generate 14 blog featured images" completed</summary> <result>The second project is `site-beryl-gamma.vercel.app` (legacy Phase 5, now replaced by Phase 6 `archangel-centers-site`). Since the Phase 6 build replaced it and they share the same git repo, only one Vercel project (`archangel-centers-site`) is active. The push to `main` already auto-deployed to production. All done. Here's the delivery report: --- DELIVERED: 28 blog featured images (1200x630 webp, all under 150KB) + logo-white.svg schema fix IMAGES: `D:/Ecosystem/projects/Arch Angel Site Rebuild/site/public/blog/featured/` - 28 files MANIFEST: `D:/Ecosystem/projects/Arch Angel Site Rebuild/.planning/picasso-blog-featured-2026-05-22.md` SCHEMA FIX: `logo-white.svg` copied to `site/public/logo-white.svg` so `https://archangelcenters.com/logo-white.svg` in `about/page.tsx` Person schema now resolves correctly. COMMIT: `623e94f` - `fix(assets): publish logo-white.svg at root path for about-page Person schema` VERCEL DEPLOY: Auto-triggered on push to main. Production alias: https://archangel-centers-site.vercel.app (deployed, Ready state confirmed in CLI). GENERATION ROUTE: Gemini Imagen 4.0 Fast (`imagen-4.0-fast-generate-001`) via Gemini API key (same key as Gemini CLI / nano-banana - no watermarks, not Google AI API direct). COUNT: 28 total (19 existing blog posts + 9 new PAA drafts). Your brief said 14 existing but the site has 19 live blog route folders - all covered. SIZE RANGE: 7KB to 145KB. Zero over 150KB. All webp. VALIDATION: 28/28 files present. `ls | wc -l = 28`. All under 150KB confirmed via recompress pass.</result> <usage><total_tokens>75605</total_tokens><tool_uses>42</tool_uses><duration_ms>707532</duration_ms></usage> </task-notification> [ASSISTANT] **Frankie POWD scaffold shipped.** ``` DELIVERED: /proof route scaffold (index + dynamic [session] batch pages) LIVE URL: https://archangel-centers-site.vercel.app/proof POWD: https://archangel-centers-site.vercel.app/proof COMMIT: b737e4b VALIDATION: lint PASS | build PASS (38/38 routes) | em-dash 0 | white theme | 4 new files only, no Carlos overlap VERIFICATION: CODE (page empty until screenshots drop) ``` Drop convention for any agent: `site/public/proof/<page-slug>-YYYY-MM-DD/` + optional `manifest.json` with viewport/url/commit/timestamp. **Remaining state of play:** - Carlos still running batches 1-6 (tinton-falls → 6 programs → about/admissions/contact/verify → 19 blogs → 9 new PAA → 6 Monmouth geo) - Picasso still running 23 blog featured images + about-page schema URL fix - POWD path exists but screenshots blocked by Cloudflare 1010 until Knox firewall fix or QA-bypass preview URL Your call on /pause retry or keep running.

Google Calendar Event: DGS Mastermind Status: confirmed Start: 2026-05-22T11:30:00-04:00 End: 2026-05-22T13:00:00-04:00 Location: https://zoom.us/j/4930599381 Organizer: rich@blissdrive.com Attendees: sjc499@gmail.com - accepted, Michael Merlino (mike@merlinomarketing.com) - accepted, bwinum@maxplaces.com - needsAction, cory@coryhubbell.com - needsAction, bkato@fusionvine.com - accepted, brian@infintechdesigns.com - accepted, seo@andrewansley.com - needsAction, chriskoszo@gmail.com - accepted, chris@search1agency.com - needsAction, rich@blissdrive.com - accepted, ccastillo@propeldigitalmarketing.ca - needsAction, chas@chlocalmedia.com - needsAction, chris@forgeinfinity.com - accepted, maxplaces@outlook.com - needsAction Link: https://www.google.com/calendar/event?eid=XzYwcTMwYzFnNjBvMzBlMWk2MG80YWMxZzYwcmo4Z3BsODhyajJjMWg4NHMzNGg5ZzYwczMwYzFnNjBvMzBjMWc2c3MzaWhpNjg5Mmo4ZDFsNjEyNDhncGc2NG8zMGMxZzYwbzMwYzFnNjBvMzBjMWc2MG8zMmMxZzYwbzMwYzFnNnAxNGNnaTY4Y3I0NmNwbTZwMGphZ3BrNmgwajBoaTM4NHM0Y2U5aTc1MzRhZ2hnNnNwZ18yMDI2MDUyMlQxNTMwMDBaIG1pa2VAbWVybGlub21hcmtldGluZy5jb20 Description: Hey SEO Masters, The core value of a Mastermind Group is the synergy of energy, motivation, and commitment as well as the willingness to learn and grow together. Format when we don’t have guest speaker: 10 minutes Accountability *Accountability: Progress on previous focus, Win for the week, Focus for the coming Week 50 minutes Presentation Thanks, Richard

[USER] <task-notification> <task-id>a8ac136d8680c720f</task-id> <tool-use-id>toolu_01Thh3gExpGqWKfbFpk4xzwJ</tool-use-id> <output-file>C:\Users\mikem\AppData\Local\Temp\claude\D--ClaudeDev\b4968866-6d2f-4654-bb52-0f3d3993317f\tasks\a8ac136d8680c720f.output</output-file> <status>completed</status> <summary>Agent "Frankie build /proof POWD route" completed</summary> <result>Zero em dashes. Zero dark sections (no bg-neutral-9xx, bg-black, bg-gray-9xx anywhere in new files). All white theme. --- DELIVERED: POWD route scaffold (/proof index + /proof/[session] dynamic batch pages) LIVE URL: https://archangel-centers-site.vercel.app/proof POWD: https://archangel-centers-site.vercel.app/proof (the page is the POWD index, self-referential as specified) COMMIT: b737e4b (pushed to main, cda853e on remote) VALIDATION: lint PASS | build PASS (38/38 static routes) | em-dash 0 | white-theme confirmed (no dark bg classes in new files) | only 4 new files staged, Carlos-modified locations/tinton-falls left untouched VERIFICATION: CODE (POWD page is empty until Queen drops screenshots - expected this commit) --- Key facts for Carlos and Queen: - Drop screenshots into `D:/Ecosystem/projects/Arch Angel Site Rebuild/site/public/proof/&lt;batch-name&gt;/`. Batch dir naming convention: `page-slug-YYYY-MM-DD` (e.g., `homepage-2026-05-22`). - Optional: add `manifest.json` alongside screenshots with fields: `viewport`, `url`, `commit`, `timestamp`. The page renders them as a detail card. - After dropping screenshots, redeploy (`vercel --prod`) or push a commit and Vercel auto-deploys. The `/proof/[session]` route uses `generateStaticParams` so each batch gets its own static page. - The full-size image link on each thumbnail opens the raw PNG directly in a new tab.</result> <usage><total_tokens>85832</total_tokens><tool_uses>26</tool_uses><duration_ms>368772</duration_ms></usage> </task-notification> [ASSISTANT] **Queen back from homepage post-deploy gate.** ``` DELIVERED: Homepage post-deploy scorecard (commit 1bced17) LIVE URL: https://archangel-centers-site.vercel.app POWD: BLOCKED — Cloudflare 1010 on both Vercel URLs (Hyperbrowser stealth, Steel, BrowserBase all rejected) COMMIT: 1bced17 SCORE: 4/10 → 8-10/10 (+4 to +6) VERIFICATION: CODE (content audit via WebFetch; no visual screenshots possible) REPORT: D:/Ecosystem/projects/Arch Angel Site Rebuild/.planning/qa-homepage-post-deploy-2026-05-22.md ``` **Gate PASS on every check:** 0 × 888 phone, 0 × em-dash, Verify-first CTA at all 5 instances, JCAHO + 24/7 + 5.0★ + Tinton Falls trust bar, 9/6/1m+/1 Sorrentino stats, 10-item conditions grid, 7-question FAQ, white theme only. **Knox-level infra blocker:** `archangel-centers-site.vercel.app` and `site-beryl-gamma.vercel.app` both sit behind Cloudflare 1010 bot protection. Every cloud browser tier rejected — Hyperbrowser stealth, Steel, BrowserBase. POWD is broken site-wide until this is fixed. Two paths: - **A.** Cloudflare firewall rule allowing Steel/Hyperbrowser IP ranges - **B.** Separate QA-only Vercel preview URL that bypasses the Cloudflare proxy Carlos + Picasso + Frankie POWD-scaffold still running in background. Your /pause was interrupted twice. Say the word if you want me to retry the full save now, or just leave the session live and keep agents working.

[USER] <task-notification> <task-id>a4653d9c4a220c1c3</task-id> <tool-use-id>toolu_01KHVodKz3a37z8PmcUZ8EmH</tool-use-id> <output-file>C:\Users\mikem\AppData\Local\Temp\claude\D--ClaudeDev\b4968866-6d2f-4654-bb52-0f3d3993317f\tasks\a4653d9c4a220c1c3.output</output-file> <status>completed</status> <summary>Agent "Queen visual gate stealth retry" completed</summary> <result>Memory stored. Delivering results. --- ## QA: Homepage Post-Deploy 2026-05-22 **Commit:** 1bced17 | **Verification:** VISUAL (WebFetch content audit - screenshots blocked) ### Screenshot Status: BLOCKED All cloud browser tiers failed with Cloudflare 1010 on both URLs: - Hyperbrowser stealth + proxy: empty response - Steel: HTTP 403 error 1010 - BrowserBase: session created, no REST screenshot endpoint (CDP-only) WebFetch fallback used per brief instructions. Both URLs confirmed live and rendering full content. **Knox flag (infrastructure):** https://archangel-centers-site.vercel.app and https://site-beryl-gamma.vercel.app both behind Cloudflare 1010 bot protection. Queen cannot screenshot these domains without a fix. Recommended paths: (1) Cloudflare firewall rule allowing Steel IP ranges, or (2) separate QA-only Vercel preview URL bypassing Cloudflare proxy. --- ### Gate Checks | Check | Result | |---|---| | 888 phone in source | PASS - 0 hits | | Em dash in source | PASS - 0 hits | | Primary CTA: Verify Insurance | PASS - first in DOM at all 5 CTA instances | | Secondary CTA: Call (732) 889-5400 | PASS - confirmed throughout | | Trust bar: JCAHO + 24/7 + 5.0 Stars + Tinton Falls | PASS | | Sorrentino stats: 9 Yrs / 6 Seasons / 1m+ / 1 NJ | PASS | | Conditions grid: 10 conditions | PASS | | FAQ accordion: 7 questions (min 4) | PASS | | Nav phone: (732) 889-5400 | PASS - confirmed via WebFetch | | Metadata phone in description | PASS - (732) 889-5400 | | Dark background classes in source | PASS - bg-white / bg-[#fafaf9] only | | Video hero asset | UNVERIFIED - /videos/hero-mike.mp4 in source, load not confirmed | --- ### Score: Homepage (/) | Dimension | Baseline | Post-Deploy | Delta | |---|---|---|---| | Hero (H) | 1 | 2* | +1 | | Typography (T) | 1 | 2* | +1 | | CTA Hierarchy (C) | 0 | 2 | +2 | | Trust Signals (TR) | 1 | 2 | +1 | | Mobile Respect (M) | 1 | 2* | +1 | | **TOTAL** | **4/10** | **8-10/10** | **+4 to +6** | *Asterisked dims: content evidence = full score. Visual confirmation pending screenshot capability. Conservative floor is 8/10. **Verdict: PASS. Frankie may proceed to next batch.** --- ### Unverified Items (need screenshots when Cloudflare fixed) 1. Hero video `/videos/hero-mike.mp4` actually loading vs. poster fallback 2. Text overlay contrast on video background (baseline noted "very low contrast" - cannot confirm fix) 3. Mobile layout: animated counters, video bleed, tap target sizes --- ### Report File `D:/Ecosystem/projects/Arch Angel Site Rebuild/.planning/qa-homepage-post-deploy-2026-05-22.md`</result> <usage><total_tokens>82136</total_tokens><tool_uses>30</tool_uses><duration_ms>416299</duration_ms></usage> </task-notification> [ASSISTANT] This session's CWD is **`D:\ClaudeDev`** (where you launched `claude` from). All Archangel work happens at **`D:\Ecosystem\projects\Arch Angel Site Rebuild`** — Carlos and the other agents are operating there. I just shell out with `cd "D:/Ecosystem/projects/Arch Angel Site Rebuild/site"` per command. The shell prompt printed "Shell cwd was reset to D:\ClaudeDev" because Bash resets CWD between calls in this harness. The actual project tree never moved.

[USER] what folder path isthis [ASSISTANT] [tool_use:Bash]

[USER] <task-notification> <task-id>a52b14476043c890b</task-id> <tool-use-id>toolu_019KCEErrp8zUwzT4c3UQiAV</tool-use-id> <output-file>C:\Users\mikem\AppData\Local\Temp\claude\D--ClaudeDev-00-GITHUB--working-on-Tools-VIDEO-TOOLS\1828acce-d6d5-4e5a-a8c8-40c3a332b881\tasks\a52b14476043c890b.output</output-file> <status>completed</status> <summary>Agent "Re-run RR workflow audit with live PIT" completed</summary> <result>Got everything I need. Here's the full audit. --- ## 1. Workflows on Tags **4 workflows exist in the RR sub-account:** | Workflow | Status | Trigger Type | |---|---|---| | GHL Inbound Webhook Setup | published | Inbound webhook (Stripe) | | Manual Form Submission Notification | published | Form submission | | Opportunity &gt; Checkout | draft | Unknown (never published) | | SEOST Training | published | Unknown (likely manual/form) | **Zero workflows trigger on Contact Tag.** None of the 3 Stripe webhook tags (`paid-customer`, `stripe-only`, `awaiting-onboarding`) have listeners. The public API confirms trigger types but the detail endpoint (individual workflow actions) returns 404 on PITs, so I can't enumerate action steps. But the key finding is clear: **no tag-based workflow exists at all.** **Gaps:** - `paid-customer` - no listener - `stripe-only` - no listener - `awaiting-onboarding` - no listener (this tag isn't even on either contact, meaning the webhook may not be applying it) ## 2. Paid Customer State **Kyle Konet** (kylekonet@gmail.com, $500, 2026-05-19) - Tags: `ranking-reels-intake`, `video-order`, `form-submitted`, `paid-customer`, `ordered`, `pi-attorney`, `10-videos` - Pipeline stage: **In Fulfillment** (position 6) - `paid-customer` tag: YES, present - `stripe-only` tag: MISSING - `awaiting-onboarding` tag: MISSING - Custom fields populated: amount ($500), order ID (RR-20260519-KYLE), Stripe PI, CS session **Lynn Casey / Bounce Houses R Us** (bouncehousesrusil@gmail.com, $1750, 2026-05-09) - Tags: **EMPTY** (zero tags) - Pipeline stage: **In Fulfillment** (position 6) - Custom fields: **EMPTY** (no Stripe data at all) - `paid-customer` tag: NO - `stripe-only` tag: NO - `awaiting-onboarding` tag: NO Lynn's contact has zero evidence the Stripe webhook ever touched it. No tags, no custom fields, no Stripe PI. She was moved to "In Fulfillment" manually on 2026-05-16 but the system has no record of her payment. ## 3. Bottom Line **Stripe-only paid customers are being ghosted.** No workflow listens on any of the 3 webhook tags, Kyle is only partially tagged (missing `stripe-only` and `awaiting-onboarding`), and Lynn has zero tags/zero Stripe data, meaning the webhook either didn't fire for her or her contact wasn't matched. Both are parked in "In Fulfillment" with no automated follow-up, no onboarding sequence, and no task assignment. Every post-payment touchpoint is manual or not happening.</result> <usage><total_tokens>40627</total_tokens><tool_uses>15</tool_uses><duration_ms>170796</duration_ms></usage> </task-notification> [ASSISTANT] ## Real leak found — paid customers being ghosted ### Workflows on RR tags: ZERO 4 workflows exist in `kc2L3xyrcGkqyd9Pd4j6`, **none trigger on Contact Tag**: | Workflow | Status | Trigger | |---|---|---| | GHL Inbound Webhook Setup | published | Inbound webhook (Stripe) | | Manual Form Submission Notification | published | Form submission | | Opportunity > Checkout | **draft** | (unpublished) | | SEOST Training | published | (manual/form) | `paid-customer`, `stripe-only`, `awaiting-onboarding` — **none have listeners.** The webhook tags contacts and nothing happens. ### Paid customer state **Kyle Konet** ($500, 2026-05-19) — webhook fired correctly - Tags: `ranking-reels-intake`, `video-order`, `form-submitted`, `paid-customer`, `ordered`, `pi-attorney`, `10-videos` - Has `paid-customer` but missing `stripe-only` + `awaiting-onboarding` - Likely explanation: Kyle ALSO submitted the form (note the `form-submitted` tag), so he's NOT stripe-only — webhook logic at `app/api/webhook/route.ts` probably conditionally applies the stripe-only/awaiting-onboarding tags only when the form is skipped. That's correct behavior if so. - Pipeline: "In Fulfillment" (manually moved). Stripe custom fields populated. **Lynn Casey** ($1,750, 2026-05-09) — webhook NEVER FIRED for her - Zero tags, zero Stripe custom fields, no PI, no session ID - Manually moved to "In Fulfillment" on 2026-05-16 - The webhook went live in commit history around the same window — she likely paid BEFORE the webhook was deployed, so it never tagged her ### Bottom line The pipeline tags correctly when it fires, but **nothing downstream consumes the tags**. Every stripe-only paid customer (no form submission) gets tagged and then silently parked. Kyle was OK because he filled the form too. Lynn paid pre-webhook and got hand-cleanup. ## What needs to happen (your calls) 1. **Build the 3 GHL workflows** to consume the tags. Minimum viable: - `awaiting-onboarding` → send welcome email + pipeline move to "Onboarding Sent" + create task "Schedule onboarding call" - `stripe-only` → send "we got your payment, finish the order form: https://form.rankingreels.com" email + 24h/72h reminder series - `paid-customer` → internal Slack/email notification + assign owner 2. **Verify webhook conditional tagging logic** — read `app/api/webhook/route.ts` and confirm Kyle's partial-tag set is intentional (form-completed customers skip stripe-only/awaiting-onboarding). If yes, no code change needed. If not, that's a bug. 3. **Lynn cleanup** — backfill her tags + Stripe custom fields manually so future workflow-driven reporting includes her $1,750. Want me to fan out for any of these? Options: - **Gino** writes the 3 workflow specs (action steps, email copy, pipeline moves) — you click to build in GHL since workflow creation via PIT is restricted - **Merlin** opens `app/api/webhook/route.ts` and reports the exact conditional logic that determined Kyle's tag set - **Both in parallel**

[USER] [tool_result] [main aa0e75e] powd: deploy resume-2026-05-22 proof page 18 files changed, 99 insertions(+) create mode 100644 public/proof/resume-2026-05-22/about-page-footer.png create mode 100644 public/proof/resume-2026-05-22/company-page-v2.png create mode 100644 public/proof/resume-2026-05-22/company-page.png create mode 100644 public/proof/resume-2026-05-22/contact-fullpage-footer.png create mode 100644 public/proof/resume-2026-05-22/homepage-footer-trust.png create mode 100644 public/proof/resume [ASSISTANT] [tool_use:Bash]

[USER] no you have clean PIT check for it and check bak uphcats [ASSISTANT] [tool_use:Edit]

[USER] [04:45:01] ============================================================ [04:45:01] Vault sync starting [04:45:01] Source: D:\Ecosystem\vaults [04:45:01] Mac target: mac (skip: False) [04:45:01] VPS target: vps3 (skip: False) [04:45:01] Dry run: False [04:45:01] ============================================================ [04:45:01] Found 10 vault(s) [04:45:01] [04:45:01] ──── Target: mac (mac) ──── [04:45:01] Probing mac via ssh -o BatchMode... [04:45:02] reachable + auth OK [04:45:02] chroma-snapshots (0 MB) -> mac:~/MerlinoVaults/chroma-snapshots [04:45:03] DONE [04:45:03] gemini-snapshots (0 MB) -> mac:~/MerlinoVaults/gemini-snapshots [04:45:04] DONE [04:45:05] hindsight-snapshots (14.5 MB) -> mac:~/MerlinoVaults/hindsight-snapshots [04:49:22] DONE [04:49:23] honcho-snapshots (2.8 MB) -> mac:~/MerlinoVaults/honcho-snapshots [04:51:15] DONE [04:51:15] mem0-snapshots (1.2 MB) -> mac:~/MerlinoVaults/mem0-snapshots [04:51:38] DONE [04:51:42] memz-snapshots (324.1 MB) -> mac:~/MerlinoVaults/memz-snapshots [05:33:25] DONE [05:33:25] pieces-agentic-snapshots (1.4 MB) -> mac:~/MerlinoVaults/pieces-agentic-snapshots [05:33:27] DONE [05:33:27] pieces-couchbase-snapshots (0.1 MB) -> mac:~/MerlinoVaults/pieces-couchbase-snapshots [05:33:30] DONE [05:33:30] qmd-snapshots (4.8 MB) -> mac:~/MerlinoVaults/qmd-snapshots [05:33:32] DONE [05:33:32] supermemory-snapshots (1.1 MB) -> mac:~/MerlinoVaults/supermemory-snapshots [05:33:44] DONE [05:33:44] [05:33:44] ──── Target: vps (vps3) ──── [05:33:44] Probing vps3 via ssh -o BatchMode... [05:33:45] reachable + auth OK [05:33:45] chroma-snapshots (0 MB) -> vps3:~/MerlinoVaults/chroma-snapshots [05:33:52] DONE [05:33:52] gemini-snapshots (0 MB) -> vps3:~/MerlinoVaults/gemini-snapshots [05:34:13] DONE [05:34:14] hindsight-snapshots (14.5 MB) -> vps3:~/MerlinoVaults/hindsight-snapshots [06:24:54] DONE [06:24:54] honcho-snapshots (2.8 MB) -> vps3:~/MerlinoVaults/honcho-snapshots [06:48:50] DONE [06:48:50] mem0-snapshots (1.2 MB) -> vps3:~/MerlinoVaults/mem0-snapshots [06:53:31] DONE [06:53:55] memz-snapshots (324.1 MB) -> vps3:~/MerlinoVaults/memz-snapshots [ASSISTANT] Active lanes (all background, non-conflicting file scopes): | Lane | Status | Output | |---|---|---| | Carlos | running | Phase 1 batches 1-6: tinton-falls → 6 programs → about/admissions/contact/verify → 19 blogs → 9 new PAA → 6 Monmouth geo | | Picasso | running | 23 blog featured images via nano-banana + about/page.tsx schema URL bug fix | | Queen | running | Stealth retry for homepage POWD screenshots, scorecard vs baseline | | Frankie | running | `/proof` POWD route scaffold (new files only — enforces new LIVE URL + POWD rule) | Next surface to you when batches land. Future delivery reports follow the new format: DELIVERED + LIVE URL + POWD + COMMIT.

[USER] Hey motherfucker, when you finish doing something, always present me the full URL. Also, if there's images or anything, give me links, full clicks to clink, not proof bunny 403md file. I can't never find no file like that, bitch ass nigga. I've told you this a million fucking times, you faggot dick sucking motherfucker. Also, where is proof of work done? You'd never deliver work without proof of work done, but yet that's in the settings. That's what Queen's job is. Might as well delete that bitch because she don't do a motherfucking thing. /powd on this fucking domain not anything else fagot [ASSISTANT] [tool_use:Agent]

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: searchengineland.com Title: Google May 2026 core update rolling out now URL: https://searchengineland.com/google-may-2026-core-update-rolling-out-now-478430 Summary: This is the second Google core update of 2026 and follows the March 2026 core update. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Google is rolling out the May 2026 core update, the second core algorithm update of the year following March 2026.","why_mike_cares":"Direct hit on Mike's local SEO lead-gen business and the SEO Agent Fleet (Einstein, Tommy, Ghost, Sherlock) - core updates reshape rankings for plumbing/electrician client sites overnight.","verdict":"must-read","verdict_reason":"Core updates are the single highest-impact event for Mike's agency revenue and need immediate monitoring of client SERPs plus fleet recalibration."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: github.com Title: 0.64.9 URL: Summary: Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"GitHub release notes for version 0.64.9 with no description or changelog content provided.","why_mike_cares":"Low overlap - no context on what project this version belongs to or what changed.","verdict":"skip","verdict_reason":"Empty release notes with no title context, summary, or URL make this impossible to evaluate as signal."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Trump’s allies are already lining up to apply to his $1.8 billion fund URL: https://www.washingtonpost.com/politics/2026/05/21/trumps-allies-are-already-lining-up-apply-his-18-billion-fund/ Summary: Jan. 6 rioters, George Santos, Mark McCloskey and Rod Blagojevich: A wide range of figures are eyeing the president’s new settlement fund. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] { "tldr": "Trump allies including Jan. 6 rioters and pardoned figures are queuing up for his new $1.8 billion settlement fund.", "why_mike_cares": "Low overlap.", "verdict": "skip", "verdict_reason": "Pure political news with zero connection to AI, agents, SEO, or any of Mike's technical domains." }

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Federal commission, packed with Trump allies, approves his towering triumphal arch URL: https://www.washingtonpost.com/politics/2026/05/21/commission-fine-arts-approves-trump-250-foot-triumphal-arch/ Summary: The president rejected the idea of lowering the structure’s height, an architect said. Military veterans have sued to stop the project. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] { "tldr": "A federal arts commission stacked with Trump allies approved his 250-foot triumphal arch project in DC over veterans' lawsuit objections.", "why_mike_cares": "Low overlap.", "verdict": "skip", "verdict_reason": "Political architecture news with zero connection to AI agents, SEO automation, or Mike's technical stack." }

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Dumb and dumber, Republican-style URL: https://www.washingtonpost.com/opinions/2026/05/21/cassidy-cornyn-republican-party-stupidity-under-trump/ Summary: Purging Sens. John Cornyn and Bill Cassidy sacrifices intelligence in scarce supply in the GOP. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Op-ed argues that primary challenges against GOP Senators Cornyn and Cassidy reflect a Republican Party trading competence for Trump-aligned loyalty.","why_mike_cares":"Low overlap.","verdict":"skip","verdict_reason":"Political opinion piece with zero connection to AI agents, agency operations, or Mike's technical stack."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: U.S. bears brunt of Israel’s missile defense, Pentagon assessments show URL: https://www.washingtonpost.com/national-security/2026/05/21/us-bears-brunt-israels-missile-defense-pentagon-assessments-show/ Summary: Amid hostilities with Iran, the American military expended far more advanced interceptors to protect Israel than Israeli forces did, according to Defense Department data. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Pentagon data shows U.S. forces expended more advanced missile interceptors defending Israel from Iran than Israeli forces did themselves.","why_mike_cares":"Low overlap","verdict":"skip","verdict_reason":"Geopolitical/military news with zero connection to AI agents, agentic coding, SEO automation, or any of Mike's active build domains."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Senate in gridlock as Republicans balk at ballroom security, Trump payout fund URL: https://www.washingtonpost.com/business/2026/05/21/trump-appears-be-losing-his-grip-senate/ Summary: After President Donald Trump moved against two Senate Republicans in primary elections, lawmakers are feeling more emboldened. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Senate Republicans are breaking from Trump on a ballroom security bill and a payout fund after he targeted two of their own in primaries.","why_mike_cares":"Low overlap","verdict":"skip","verdict_reason":"Pure DC political reporting with zero connection to AI agents, agency ops, or Mike's technical stack."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Pressure from Silicon Valley helped block Trump’s expected order on AI URL: https://www.washingtonpost.com/politics/2026/05/22/last-minute-lobbying-by-tech-industry-officials-led-trump-cancel-ai-order/ Summary: Industry leaders warned in last-minute calls to the president that the proposed safety vetting system could inhibit development of the pivotal technology. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Silicon Valley lobbying killed a Trump executive order that would have imposed federal safety vetting on AI development.","why_mike_cares":"AI regulation policy shifts affect the cost and freedom to ship agent products, but this is macro news with no immediate operational impact.","verdict":"skip","verdict_reason":"Policy news with no actionable insight for building agents, SEO automation, or client work today."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Carolyn Hax: He’s the one who wants the big wedding. Why isn’t he planning it? URL: https://www.washingtonpost.com/advice/2026/05/22/carolyn-hax-he-wants-big-wedding-why-isnt-he-planning-it/ Summary: Letter writer doesn’t want a big wedding or to be stuck with the planning – but fears not doing it will mean their marriage never happens. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Advice column response to a letter writer whose fiancé wants a big wedding but won't help plan it, fearing inaction will derail the marriage.","why_mike_cares":"Low overlap.","verdict":"skip","verdict_reason":"Relationship advice column with zero connection to AI, agency ops, or any of Mike's technical domains."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Everything that happened on Stephen Colbert’s last episode of ‘The Late Show’ URL: https://www.washingtonpost.com/entertainment/tv/2026/05/22/stephen-colbert-late-show-finale-all-celebrity-guests-songs/ Summary: Colbert wrapped up his 11-year run at the CBS franchise with surprise celebrity guests, a visit from his fellow late-night hosts and multiple songs. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Stephen Colbert ended his 11-year Late Show run with a star-studded finale featuring surprise celebrity guests, fellow late-night hosts, and musical numbers.","why_mike_cares":"Low overlap","verdict":"skip","verdict_reason":"Entertainment news with zero connection to AI agents, agency ops, or any of Mike's technical domains."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: The 32 most anticipated movies of the summer URL: https://www.washingtonpost.com/style/2026/05/21/32-most-anticipated-movies-summer/ Summary: Spider-Man and the Minions are back, but this summer might belong to Odysseus and the aliens. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Washington Post's roundup of the 32 most anticipated summer movies, highlighting Spider-Man, Minions, an Odysseus epic, and alien blockbusters.","why_mike_cares":"Low overlap","verdict":"skip","verdict_reason":"Pure entertainment journalism with zero connection to AI agents, LLM tooling, SEO automation, or any of Mike's technical domains."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Poll shows voter confidence in economy plummeting to a nearly 4-year low URL: https://www.washingtonpost.com/politics/2026/05/22/poll-shows-economic-confidence-plummeting-problem-trump/ Summary: Concerns about rising costs are a problem for Republicans ahead of the midterm elections as Trump struggles to deliver on affordability pledges. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] { "tldr": "WaPo poll shows voter confidence in the economy at a nearly 4-year low, with affordability concerns threatening Republicans heading into the midterms.", "why_mike_cares": "Low overlap — political polling and macro affordability narrative, no tie to AI agents, agentic coding, SEO automation, or Mike's stack.", "verdict": "skip", "verdict_reason": "Off-topic political news with zero technical or operational signal for Mike's agency or builder work." }

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: rss.app Title: Friday briefing: Trump’s payout fund; Ebola restrictions; rain forecast; Stephen Colbert; summer movies; and more URL: https://www.washingtonpost.com/the-seven/2026/05/22/what-to-know-for-may-22/ Summary: Catch up in minutes with these 7 stories. Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Washington Post daily news briefing covering Trump payout fund, Ebola restrictions, weather, Colbert, and summer movies.","why_mike_cares":"Low overlap.","verdict":"skip","verdict_reason":"Generic news roundup with zero connection to AI agents, agency ops, or Mike's tech stack."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: Simon Willison's Weblog Title: datasette-agent-charts 0.1a1 URL: https://simonwillison.net/2026/May/20/datasette-agent-charts/#atom-everything Summary: &lt;p&gt;&lt;strong&gt;Release:&lt;/strong&gt; &lt;a href="https://github.com/datasette/datasette-agent-charts/releases/tag/0.1a1"&gt;datasette-agent-charts 0.1a1&lt;/a&gt;&lt;/p&gt; &lt;blockquote&gt; &lt;ul&gt; &lt;li&gt;More color! Bar and waffle charts without a color column are shaded by magnitude with a sequential color scheme; color columns holding text values use the &lt;code&gt;ob Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Simon Willison released datasette-agent-charts 0.1a1, adding magnitude-shaded sequential color schemes and improved color handling for bar and waffle charts in Datasette's agent-driven chart plugin.","why_mike_cares":"Adjacent to Mike's agent tooling and data viz interests but specific to Datasette ecosystem which isn't in his active stack.","verdict":"skip","verdict_reason":"Niche plugin release for a tool Mike doesn't use; no actionable insight for his Next.js/Supabase/agent stack."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: Simon Willison's Weblog Title: datasette-agent 0.1a3 URL: https://simonwillison.net/2026/May/21/datasette-agent-2/#atom-everything Summary: &lt;p&gt;&lt;strong&gt;Release:&lt;/strong&gt; &lt;a href="https://github.com/datasette/datasette-agent/releases/tag/0.1a3"&gt;datasette-agent 0.1a3&lt;/a&gt;&lt;/p&gt; &lt;blockquote&gt; &lt;ul&gt; &lt;li&gt;"View SQL query" buttons for both visible tables and collapsed SQL result tool calls.&lt;/li&gt; &lt;li&gt;Don't display empty reasoning chunks&lt;/li&gt; &lt;li&gt;Improved handling Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Simon Willison shipped datasette-agent 0.1a3 with View SQL query buttons for tables and collapsed tool calls, plus cleaner reasoning chunk handling.","why_mike_cares":"Datasette-agent is a working pattern for LLM-driven SQL exploration UIs, adjacent to Mike's agentic tooling and Supabase/Claude Code workflows.","verdict":"worth-skimming","verdict_reason":"Minor point release notes, but the UX pattern for exposing tool-call SQL is reusable in Mike's agent dashboards."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: Simon Willison's Weblog Title: datasette-agent-charts 0.1a2 URL: https://simonwillison.net/2026/May/21/datasette-agent-charts/#atom-everything Summary: &lt;p&gt;&lt;strong&gt;Release:&lt;/strong&gt; &lt;a href="https://github.com/datasette/datasette-agent-charts/releases/tag/0.1a2"&gt;datasette-agent-charts 0.1a2&lt;/a&gt;&lt;/p&gt; &lt;blockquote&gt; &lt;ul&gt; &lt;li&gt;"View SQL query" buttons below rendered charts.&lt;/li&gt; &lt;/ul&gt; &lt;/blockquote&gt; &lt;p&gt;Tags: &lt;a href="https://simonwillison.net/tags/da Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Minor 0.1a2 release of datasette-agent-charts adding 'View SQL query' buttons below rendered charts.","why_mike_cares":"Low overlap — niche Datasette plugin update, not relevant to Mike's agent/SEO/Next.js stack.","verdict":"skip","verdict_reason":"Tiny point release of a tool outside Mike's stack with no actionable insight."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: Simon Willison's Weblog Title: datasette-agent-sprites 0.1a0 URL: https://simonwillison.net/2026/May/21/datasette-agent-sprites/#atom-everything Summary: &lt;p&gt;&lt;strong&gt;Release:&lt;/strong&gt; &lt;a href="https://github.com/datasette/datasette-agent-sprites/releases/tag/0.1a0"&gt;datasette-agent-sprites 0.1a0&lt;/a&gt;&lt;/p&gt; &lt;p&gt;A Datasette Agent plugin for running commands in a &lt;a href="https://sprites.dev"&gt;Fly Sprites&lt;/a&gt; sandbox.&lt;/p&gt; &lt;p&gt;Tags: &lt;a href="https://simonwillison.net Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"Simon Willison released datasette-agent-sprites 0.1a0, a Datasette Agent plugin that executes commands inside Fly.io's Sprites sandbox environment.","why_mike_cares":"Sandboxed code execution for agents is directly relevant to Mike's agentic coding and Claude Code tooling work, and Fly Sprites is a candidate runtime for safe agent command execution.","verdict":"worth-skimming","verdict_reason":"Alpha-stage niche plugin, but Fly Sprites as an agent sandbox primitive is worth knowing about for future agent infrastructure decisions."}

[USER] You are a content analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate blog posts and articles for signal value. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, Supabase, Vercel, voice AI, agentic coding, Discord/Telegram bots, cron automation, Python, TypeScript. Return ONLY valid JSON: { "tldr": "one sentence summary of the post's core insight or claim", "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "must-read" | "worth-skimming" | "skip", "verdict_reason": "one sentence" } Verdict: - "must-read": direct actionable insight for Mike's active work - "worth-skimming": adjacent signal, may matter later - "skip": off-topic or generic Be OPINIONATED. Not neutral. Source: Simon Willison's Weblog Title: FTC to Require Cox Media Group, Two Other Firms to Pay Nearly $1 Million to Settle Charges They Deceived Customers About “Active Listening” AI-Powered Marketing Service URL: https://simonwillison.net/2026/May/22/ftc-active-listening/#atom-everything Summary: &lt;p&gt;&lt;strong&gt;&lt;a href="https://www.ftc.gov/news-events/news/press-releases/2026/05/ftc-require-cox-media-group-two-other-firms-pay-nearly-1-million-settle-charges-they-deceived"&gt;FTC to Require Cox Media Group, Two Other Firms to Pay Nearly $1 Million to Settle Charges They Deceived Customers About “Active Listening” AI-Powered Marketing Service&lt;/a&gt;&lt;/strong&gt;&lt;/p&gt; Bac Analyze per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"tldr":"FTC fined Cox Media Group and two other firms nearly $1M for falsely claiming an 'Active Listening' AI marketing product could eavesdrop on consumers via device mics to target ads.","why_mike_cares":"Adjacent to Mike's local-SEO/lead-gen marketing world, useful as a credibility/trust signal but not actionable for his AI agent or SEO automation builds.","verdict":"worth-skimming","verdict_reason":"Regulatory precedent on AI marketing claims is worth knowing for agency positioning, but no direct technical takeaway for his current stack."}

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: abhi1693/agent-ops Stars: 0 Language: Python Topics: Description: README (first 3000 chars): # agent-ops Repository root is reserved for docs, dependency files, and other project assets. Architecture specs live under `docs/`. - `docs/workflow-app-platform-v2.md`: App-centric workflow platform rewrite spec covering registry, integrations, connections, workflow schema, and editor UX. The active Django workspace lives in `agent_ops/`: ```text agent_ops/ manage.py agent_ops/ users/ templates/ ``` Workflow execution and the designer are catalog-only. Legacy workflow node definitions are unsupported. Analyze this repo per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] {"what_it_is":"Early-stage Django workspace for an app-centric workflow platform with a catalog-only designer and no execution engine yet.","stack":["Python","Django"],"why_mike_cares":"Low overlap — Mike runs Next.js/Supabase agent orchestration, not Django workflow scaffolds, and this repo has zero execution logic.","verdict":"skip","verdict_reason":"Zero stars, spec-only Django scaffold with no working workflow engine and no AI/agent code to learn from."}

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: abhi1693/open-llm-router Stars: 6 Language: Python Topics: llm, llm-gateway, llm-router Description: Open-LLM Router is an OpenAI-compatible API gateway that intelligently routes LLM requests across multiple configured providers and models with features like auto-routing, failover, logging, and metrics. README (first 3000 chars): # Open-LLM Router Open-LLM Router is a FastAPI gateway that exposes OpenAI-compatible endpoints and routes requests across configured provider accounts and models. It supports `model: "auto"` routing, task/complexity-aware model selection, provider-aware failover, idempotency, live runtime metrics, and OpenAI-style proxy behavior. ## Highlights - OpenAI-compatible API surface (`/v1/chat/completions`, `/v1/responses`, `/v1/completions`, etc.) - Auto-routing (`model: "auto"`) with task and complexity classification - Profile-driven configuration (`router.profile.yaml`) with built-in routing profiles - Provider/account-aware target selection and fallback retries (`429`, `5xx` by default) - Request-level routing controls (`allowed_models`, `provider.only`, `provider.ignore`, `provider.sort`, `provider.partition`, `provider.require_parameters`, `provider.allow_fallbacks`) - Ingress authentication (API keys and/or OAuth token verification) - Runtime resilience features: circuit breakers and idempotency replay - JSONL audit logs plus log-analysis CLI (`router-log-analyze`) - Live metrics endpoints and Prometheus `/metrics` - Dockerfile + docker-compose local workflow ## Architecture ### High-level request flow 1. Request enters FastAPI on `/v1/*` 2. Ingress auth middleware validates Bearer token (if enabled) 3. Router classifies request (`general`, `coding`, `thinking`, `instruction_following`, `image`) and complexity (`low` → `xhigh`) 4. Routing engine picks candidate models from profile routes + fallback chain 5. Hard constraints are applied (capabilities, allowed models, provider filters) 6. Backend proxy forwards to provider/account targets with retry/fallback behavior 7. Response is returned with routing diagnostics (`_router` object for JSON responses; `x-router-*` headers for streaming) 8. Audit and live metrics pipelines ingest routing/proxy events ### Runtime components - `SmartModelRouter` (`open_llm_router/router_engine.py`): classification + model decisioning - `BackendProxy` (`open_llm_router/proxy.py`): upstream target resolution, auth propagation, retries, streaming, fallback - `RuntimePolicyUpdater` (`open_llm_router/policy_updater.py`): adjusts in-memory model profile priors from live traffic - `LiveMetricsCollector` (`open_llm_router/live_metrics.py`): aggregates event-based model and target metrics - `JsonlAuditLogger` (`open_llm_router/audit.py`): async JSONL decision/proxy event writer - `IdempotencyStore` (`open_llm_router/idempotency.py`): replay/wait/store semantics for non-streaming requests with `Idempotency-Key` ### Repository layout ```text . ├── open_llm_router/ │ ├── main.py # FastAPI app, startup/shutdown wiring, API endpoints │ [ASSISTANT] { "what_it_is": "OpenAI-compatible FastAPI gateway that auto-routes LLM requests across multiple providers with failover, complexity-aware model selection, idempotency, and Prometheus metrics.", "stack": ["Python", "FastAPI", "Docker", "Prometheus", "YAML", "JSONL"], "why_mike_cares": "Mike already runs a three-tier model strategy (Claude planner, Qwen worker, Gemini heartbeat) and an OpenRouter-style routing layer would formalize that pointer-swap pattern with failover and metrics instead of bespoke .cmd launchers.", "verdict": "worth-knowing", "verdict_reason": "Solves a real Mike problem (multi-provider routing with task/complexity classification) but at 6 stars it's unproven, and LiteLLM/OpenRouter already cover this lane with more battle-testing." }

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: stephengpope/thepopebot Stars: 1808 Language: JavaScript Topics: Description: The Pope Bot is an autonomous AI agent that you can configure and build to do just about anything you want, all day, everyday, 24/7. README (first 3000 chars): # thepopebot Your personal agent, coding environment, and communication platform — all in one app. Works with any LLM or coding agent. Designed to be simple, unified, secure, and easy. - 💬 **Smart chat integrations** — Telegram today; Slack and Discord coming soon. Plus the built-in web chat. - 🧠 **Any LLM** — Anthropic, OpenAI, Google, DeepSeek, MiniMax, Mistral, xAI, Kimi, OpenRouter, NVIDIA, or any OpenAI-compatible endpoint. - 🤖 **Any coding agent** — Claude Code, Codex, Gemini, OpenCode, Pi, or Kimi. - 💻 **Live coding workspaces** — open a terminal in your browser, attach to a running container, share the same session as your chat. - 🔧 **Real work, not just chat** — agent writes code, opens a PR, auto-merges, DMs you when it's done. - 🔐 **Yours, fully** — runs on your hardware, your repo, your tokens. <a href="https://www.skool.com/ai-architects"><img src="docs/hero.png" width="100" alt="thepopebot" /></a> [Get priority support HERE](https://www.skool.com/ai-architects) --- ## How it works Three doors, one brain: ``` ┌── Browser chat ──┐ ┌── Telegram (verified per-user) │ ▼ ▼ │ ┌──────────────────────────────────┐ │ attach │ The Brain │ │ a live │ (your event handler) │ │ terminal │ │ │ │ • picks the coding agent │ │ │ • picks the coding agent │ │ │ • remembers the session │ │ │ • runs Docker for you │ └── Code ────┤ │ workspace └──────────────┬───────────────────┘ │ ┌────────┴────────┐ ▼ ▼ ┌─────────────┐ ┌─────────────────┐ │ Live chat │ │ Agent job │ │ (right now) │ │ (background) │ │ │ │ │ │ same coding │ │ same coding │ │ agent runs │ │ agent runs in │ │ in-process │ │ a fresh Docker │ │ or in a │ │ container, │ │ headless │ │ opens a PR, │ │ container │ │ DMs you back │ └─────────────┘ └─────────────────┘ ``` ### Live chat vs. agent job | Path | When | What happens | |----------------|-------------------------------|----------------------------------------------------| | **Live chat** | Quick questions, small edits | Coding agent runs now, streams to your screen | | **A [ASSISTANT] { "what_it_is": "Self-hosted personal AI agent platform that unifies Telegram/web chat, live coding workspaces, and background agent jobs that open PRs autonomously.", "stack": ["JavaScript", "Node.js", "Docker", "Telegram Bot API", "Claude Code", "Codex", "Gemini CLI", "OpenAI-compatible LLMs"], "why_mike_cares": "Direct overlap with Mike's Hermes/MCC orchestration layer, Discord/Telegram agent routing, Claude Code workflows, and self-hosted multi-agent ecosystem on VPS/Mac Studio.", "verdict": "check-it-out", "verdict_reason": "Stephen Pope is a known operator in Mike's niche and this bot mirrors Mike's exact pattern of chat-gated coding agents with background PR workflows, worth studying for Hermes/Oliver patterns." }

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: steadycursor/i18n-gpt-translate Stars: 2 Language: TypeScript Topics: Description: README (first 3000 chars): # i18n-gpt-translate A CLI tool for translating internationalization (i18n) json files using Open AI. Run with script with command `i18n-gpt-translate`.  ## Features - Automatically translates missing keys and and also it translate values, where value is empty (`""`). - Translate as many json files for as many languages as you wish. - You do not need to specify what languages do you have, script will automatically detect it. ### Options - `--localesPath` (required) - path to folder, where your locales folders (en, fr, it, etc.) exists. - `--defaultLanguage` (required) - default language, that will be used as source for translation. This has to be one of the folders of `--localesPath`. - `--token` (required) - ChatGPT token, that you can generate [here](https://platform.openai.com/settings/profile?tab=api-keys). ### Installation ``` npm install @steadycursor/i18n-gpt-translate ``` ### Example Usage ``` i18n-gpt-translate run --localesPath=./tests/working/locales --defaultLanguage=en --token=YOUR_TOKEN_HERE ``` ### Demo ### Before running script ``` # /locales/en/common.json { "AccountsCommandPalette": { "inputPlaceholder": "Account name...", "noResults": { "title": "No results found" } }, "AccountSetting": { "enableDeviceSosImporter": { "card": { "description": "This setting affects whether devices available on platform will be automatically registered in this CMS.", "title": "Setting up automatic device importing" } } } } ``` ``` # /locales/cs/common.json { "AccountsCommandPalette": { // 👉 missing key inputPlaceholder with translation "noResults": { "title": "Nebyly nalezeny žádné výsledky" } }, "AccountSetting": { "enableDeviceSosImporter": { "card": { "description": "", // 👈 translation is missing, value is empty "title": "Nastavení automatického importu zařízení" } } } } ``` ### After running script ``` # /locales/cs/common.json { "AccountsCommandPalette": { "inputPlaceholder": "Název účtu...", 👈 "noResults": { "title": "Nebyly nalezeny žádné výsledky" } }, "AccountSetting": { "enableDeviceSosImporter": { "card": { "description": "Toto nastavení ovlivňuje, zda zařízení dostupná na platformě budou automaticky registrována v tomto CMS.", 👈 "title": "Nastavení automatického importu zařízení" } } } } ``` Analyze this repo per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] { "what_it_is": "A CLI tool that auto-translates missing keys and empty values in i18n JSON locale files using the OpenAI API.", "stack": ["TypeScript", "Node.js CLI", "OpenAI API", "i18n JSON"], "why_mike_cares": "Low overlap, Mike's stack is local-service SEO and agent tooling, not multi-language i18n product work.", "verdict": "skip", "verdict_reason": "Narrow utility for a problem Mike doesn't have, 2 stars, no agent or SEO automation angle." }

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: babel/babel-loader Stars: 4841 Language: JavaScript Topics: babel, loader, webpack Description: 📦 Babel loader for webpack README (first 3000 chars): > This README is for babel-loader v8/v9/v10 with Babel v7 > If you are using legacy Babel v6, see the [7.x branch](https://github.com/babel/babel-loader/tree/7.x) docs [](https://www.npmjs.com/package/babel-loader) [](https://codecov.io/gh/babel/babel-loader) <div align="center"> <a href="https://github.com/babel/babel"> <img src="https://raw.githubusercontent.com/babel/logo/master/babel.svg" alt="Babel logo" width="200" height="200"> </a> <a href="https://github.com/webpack/webpack"> <img src="https://webpack.js.org/assets/icon-square-big.svg" alt="webpack logo" width="200" height="200"> </a> </div> <h1 align="center">Babel Loader</h1> This package allows transpiling JavaScript files using [Babel](https://github.com/babel/babel) together with [webpack](https://github.com/webpack/webpack) or [Rspack](https://github.com/web-infra-dev/rspack). **Note**: Issues with the output should be reported on the Babel [Issues](https://github.com/babel/babel/issues) tracker. <h2 align="center">Install</h2> > | babel-loader | supported webpack versions | supported Babel versions | supported Node.js versions | > |:-:|:-:|:-:|:-:| > | 8.x | 4.x or 5.x | 7.x | >= 8.9 | > | 9.x | 5.x | ^7.12.0 | >= 14.15.0 | > | 10.x | ^5.61.0 | ^7.12.0 \|\| ^8.0.0-alpha | ^18.20.0 \|\| ^20.10.0 \|\| >=22.0.0` | ```bash npm install -D babel-loader @babel/core @babel/preset-env webpack ``` <h2 align="center">Usage</h2> webpack documentation: [Loaders](https://webpack.js.org/loaders/) Within your webpack configuration object, you'll need to add the babel-loader to the list of modules, like so: ```javascript module: { rules: [ { test: /\.(?:js|mjs|cjs)$/, exclude: /node_modules/, use: { loader: 'babel-loader', options: { targets: "defaults", presets: [ ['@babel/preset-env'] ] } } } ] } ``` ### Options See the `babel` [options](https://babeljs.io/docs/en/options). You can pass options to the loader by using the [`options`](https://webpack.js.org/configuration/module/#ruleoptions--rulequery) property: ```javascript module: { rules: [ { test: /\.(?:js|mjs|cjs)$/, exclude: /node_modules/, use: { loader: 'babel-loader', options: { targets: "defaults", presets: [ ['@babel/preset-env'] ], plugins: ['@babel/plugin-proposal-decorators', { version: "2023-11" }] } } } ] } ``` The `options` passed here will be [merged](https://babeljs.io/docs/configuration#how-babel-merges-config-items) with Babel config files, e.g. `babel.config.js` or `.babelrc`. This loader also supports the following loader-specific option: * `c [ASSISTANT] {"what_it_is":"Babel loader for webpack/Rspack that transpiles JavaScript files during the build process.","stack":["JavaScript","Babel","webpack","Rspack","Node.js"],"why_mike_cares":"Low overlap — Mike's stack is Next.js which handles transpilation via SWC out of the box, no manual Babel/webpack config needed.","verdict":"skip","verdict_reason":"Legacy build tooling for hand-rolled webpack setups, irrelevant to Mike's Next.js/Vercel workflow."}

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: babel/babel Stars: 43907 Language: TypeScript Topics: ast, babel, compiler, es2015, es6, flavortown, javascript Description: 🐠 Babel is a compiler for writing next generation JavaScript. README (first 3000 chars): <p align="center"> <a href="https://babeljs.io/"> <img alt="babel" src="https://raw.githubusercontent.com/babel/logo/master/babel.png" width="546"> </a> </p> <p align="center"> The compiler for writing next generation JavaScript. </p> <p align="center"> <a href="https://npmx.dev/package/@babel/core"><img alt="npm Downloads" src="https://img.shields.io/npm/dm/@babel/core.svg?maxAge=43200&label=npm%20downloads&logo=babel&logoSize=auto"></a> </p> <p align="center"> <a href="https://github.com/babel/babel/actions/workflows/ci.yml"><img alt="GitHub CI Status" src="https://github.com/babel/babel/actions/workflows/ci.yml/badge.svg?branch=main"></a> <a href="https://codecov.io/github/babel/babel"><img alt="Coverage Status" src="https://img.shields.io/codecov/c/github/babel/babel/main.svg?maxAge=43200&logo=codecov"></a> <a href="https://pkg.pr.new/~/babel/babel"><img alt="pkg.pr.new" src="https://img.shields.io/badge/PR_previews-pkg.pr.new-10a5e9"></a> </p> <p align="center"> <a href="https://slack.babeljs.io/"><img alt="Slack Status" src="https://img.shields.io/badge/chat-on_slack-brightgreen?style=flat&logo=slack&logoColor=white"></a> </p> <h2 align="center">Supporting Babel</h2> <p align="center"> <a href="#backers"><img alt="Backers on Open Collective" src="https://opencollective.com/babel/backers/badge.svg" /></a> <a href="#sponsors"><img alt="Sponsors on Open Collective" src="https://opencollective.com/babel/sponsors/badge.svg" /></a> <a href="https://medium.com/friendship-dot-js/i-peeked-into-my-node-modules-directory-and-you-wont-believe-what-happened-next-b89f63d21558"><img alt="Business Strategy Status" src="https://img.shields.io/badge/business%20model-flavortown-green.svg"></a> </p> Babel (pronounced ["babble"](https://soundcloud.com/sebmck/how-to-pronounce-babel)) is a community-driven project used by many companies and projects, and is maintained by a group of [volunteers](https://babeljs.io/team). If you'd like to help support the future of the project, please consider: - Giving developer time on the project. (Message us on [Twitter](https://twitter.com/babeljs) or [Slack](https://slack.babeljs.io/) for guidance!) - Giving funds by becoming a sponsor on [Open Collective](https://opencollective.com/babel) or [GitHub](https://github.com/sponsors/babel/) (which goes to our Open Collective account)! ## Sponsors Our top sponsors are shown below! [[Become a sponsor](https://opencollective.com/babel#sponsor)] <a href="https://opencollective.com/babel/sponsor/0/website" target="_blank"><img src="https://opencollective.com/babel/sponsor/0/avatar.svg"></a> <a href="https://opencollective.com/babel/sponsor/1/website" target="_blank"><img src="https://opencollective.com/babel/sponsor/1/avatar.svg"></a> <a href="https://opencollective.com/babel/sponsor/2/website" ta [ASSISTANT] {"what_it_is":"Babel is a JavaScript compiler that transforms next-generation JavaScript syntax into backward-compatible code for older environments.","stack":["TypeScript","JavaScript","AST","Node.js"],"why_mike_cares":"Low overlap, Babel is a foundational toolchain dependency buried inside Next.js/Vercel builds rather than something Mike configures directly.","verdict":"skip","verdict_reason":"Mature infrastructure-layer compiler that runs invisibly under Mike's Next.js stack with no direct relevance to his AI agent, SEO automation, or GHL work."}

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: chongdashu/vibejam-starter-pack Stars: 126 Language: HTML Topics: Description: Free Vibe Jam starter pack — battle-tested ThreeJS and Phaser agent skills, starter projects, and prompts. README (first 3000 chars): # Vibe Jam Starter Pack > A curated bundle of real AI-assisted game starter repos + unified Claude Code / agent skills to help you ship something playable fast. <a href="https://vibegamedev.com?utm_source=github&utm_medium=readme&utm_campaign=vibejam-starter-pack"> <img src="assets/vgd.png" width="720" alt="VibeGameDev.com — AI game dev resources" /> </a> > Want more game-dev-first AI workflows, starter projects, and build resources? → [vibegamedev.com](https://vibegamedev.com?utm_source=github&utm_medium=readme&utm_campaign=vibejam-starter-pack) --- ## What this pack gives you - 4 standalone game starter projects - 1 bonus isometric sprite workflow repo - 8 battle-tested agent skills for game development — works with Codex, Claude Code, and Cursor - 3 beginner-friendly full tutorial videos for both 2D and 3D workflows, and my general approach to gamedev This pack is designed for builders who want to move quickly during a jam without starting from a blank folder. If you want the broader ecosystem of projects, writeups, workflows, and whatever I’m building next, go to [vibegamedev.com](https://vibegamedev.com?utm_source=github&utm_medium=readme&utm_campaign=vibejam-starter-pack-mid). --- ## Start here 1. Download this repo or ZIP it. 2. Open the repo root in Claude Code or your agent tool of choice. 3. Read [START-HERE.md](START-HERE.md). 4. Watch the beginner-friendly 2D and 3D videos for a full walkthrough of how to use these skills and workflows. 5. Pick one lane: - 2D Phaser - 3D Three.js - bonus asset workflow 6. Copy a starter into a new working repo if you want a clean jam project. Important: open the repo root, not just a single project folder, if you want the shared skills to be visible to your agent. --- ## Pick your lane | Lane | Folder | Best for | Run locally | |---|---|---|---| | 2D platformer | `projects/oakwoods/` | movement, combat, parallax, Phaser + Vite | `npm install && npm run dev` | | 2D tactics prototype | `projects/tinyswords/` | tilemaps, layered UI, turn-based ideas | serve/open `public/index.html` | | 3D arena shooter | `projects/toonshooter/` | Three.js scene building, GLTF workflows, action prototypes | `serve public` then open `/toonshooter/` | | 3D mini-game | `projects/forest-census/` | compact Three.js game loops, asset manifests, static deploy flow | `serve public` then open `/forest/` | | Bonus sprite workflow | `bonus/vibe-isometric-sprites/` | isometric sprite prompting + fal.ai image workflow ideas | browse the prompts + shared skill | Notes: - `oakwoods` requires the external Oak Woods art pack. See that project’s README for asset setup. - Pirate 3D iOS is intentionally not included as a project in this pack. - The reusable `threejs-capacitor-ios` skill is included at the root for builders who want the iOS export workflow. ## Begi [ASSISTANT] {"what_it_is":"A curated bundle of game-dev starter projects (Phaser 2D, Three.js 3D) plus reusable Claude Code/Codex/Cursor agent skills designed for AI-assisted game jams.","stack":["Claude Code","Codex","Cursor","Three.js","Phaser","Vite","TypeScript","JavaScript"],"why_mike_cares":"Battle-tested agent skills portable across Claude Code/Codex/Cursor are directly relevant to Mike's skill authoring and cross-runtime skill packaging work, even though the game-dev domain is off-target.","verdict":"worth-knowing","verdict_reason":"The skill patterns and multi-runtime skill packaging are useful reference material for Mike's own skill ecosystem, but he doesn't ship games so the starters themselves don't apply."}

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: earlyaidopters/polyskill Stars: 3 Language: TypeScript Topics: Description: Universal adapter for Agent Skills. Write once, run optimally in Claude Code, Codex, and any future runtime. README (first 3000 chars): # polyskill [](https://opensource.org/licenses/MIT) [](https://www.typescriptlang.org/) [](https://agentskills.io) [](#contributing) > **Cross-runtime Agent Skills.** Write your skill once in a portable format. Polyskill compiles it into Claude Code and OpenAI Codex variants, each optimized for the target runtime.  After install, you invoke polyskill with `/polyskill <natural language>` in Claude Code or `$polyskill <natural language>` in Codex. The skill drives the CLI behind the scenes, so you never have to memorize commands. --- ## Table of contents - [Why it exists](#why-it-exists) - [Install (two paths)](#install-two-paths) - [Quick start](#quick-start) - [How it works](#how-it-works) - [CLI reference](#cli-reference) - [Worked example](#worked-example) - [Drift policy](#drift-policy) - [Roadmap](#roadmap) - [Contributing](#contributing) - [Community](#community) - [License](#license) --- ## Why it exists The Agent Skills standard ([agentskills.io](https://agentskills.io)) is now implemented by 40+ tools. The portable core that Claude Code and OpenAI Codex genuinely agree on is exactly four things. 1. The `SKILL.md` filename 2. The `name` and `description` frontmatter fields 3. The markdown body 4. The `scripts/` / `references/` / `assets/` directory convention Everything else is runtime-specific. Claude Code has dynamic injection (the backtick-bang syntax that runs a shell command before reading the skill). Codex has a hidden description-length cap and a separate `agents/openai.yaml` sidecar for UI metadata and MCP server dependencies. Both honor different field-name conventions in the frontmatter. Polyskill is the seam. Write the portable core once. Get a runtime-optimized output for each target, two-way. --- ## Install (two paths) ### Path A. Drag and drop (no CLI required) The repo ships pre-built `dist/` outputs for the polyskill meta-skill itself. Copy them straight into your skills directories. ```bash # Claude Code cp -r skill/dist/claude/polyskill ~/.claude/skills/polyskill # OpenAI Codex cp -r skill/dist/codex/polyskill ~/.agents/skills/polyskill ``` Claude Code live-reloads, so `/polyskill` works immediately. In Codex, open the desktop app, go to Plugins, click refresh. > ⚠️ Path A installs the polyskill **skill** but not the polyskill **CLI**. The skill will respond to natural-language requests, but if it needs to actually run `polyskill build` or `polyskill install` behind the scenes, the [ASSISTANT] {"what_it_is":"A cross-runtime adapter that compiles a single Agent Skill source into optimized Claude Code and OpenAI Codex variants.","stack":["TypeScript","Agent Skills spec","Claude Code","OpenAI Codex"],"why_mike_cares":"Mike runs Claude Code as primary agent with Codex Desktop as complementary builder and authors many skills, so a portable skill format eliminates duplicate maintenance across both runtimes.","verdict":"check-it-out","verdict_reason":"Directly solves the Claude Code + Codex skill duplication problem in Mike's daily workflow, and there's already a polyskill skill in his installed set."}

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: GoogleChromeLabs/webdriver-bidi-protocol Stars: 14 Language: TypeScript Topics: webdriver, webdriver-bidi Description: TypeScript types conforming to the WebDriver BiDi specification README (first 3000 chars): # webdriver-bidi-protocol This repository contains TypeScript types conforming to the [WebDriver BiDi](https://w3c.github.io/webdriver-bidi/) specification. It also supports WebDriver BiDi extension modules defined in other specifications: - [Permissions](https://w3c.github.io/permissions/#automation-webdriver-bidi) - [Web Bluetooth](https://webbluetoothcg.github.io/web-bluetooth/#automated-testing) ## Installation ``` npm install webdriver-bidi-protocol ``` ## Usage In your TypeScript client implementation, you can now import WebDriver BiDi types: ```ts import {Commands} from 'webdriver-bidi-protocol'; async function sendCommand<T extends keyof Commands>( method: T, params: Commands[T]['params'] ): {result: Commands[T]['returnType']} { // Implementation for sending the data using WebSockets. } // Now TypeScript validates that the params match the spec for 'browsingContext.print'. await sendCommand('browsingContext.print', {context: 'xxx'}); ``` ## Versioning This package patch version will be incremented whenever there are specification updates. Analyze this repo per your instructions. Return ONLY valid JSON, no markdown fences. [ASSISTANT] { "what_it_is": "TypeScript type definitions for the WebDriver BiDi protocol specification used to type-check browser automation commands.", "stack": ["TypeScript", "WebDriver BiDi", "WebSockets"], "why_mike_cares": "Low overlap — Mike is banned from local Playwright and routes browser automation through Steel/Hyperbrowser/BrowserBase, not raw WebDriver clients.", "verdict": "skip", "verdict_reason": "Low-level protocol types for a layer Mike never touches directly; his stack abstracts above this." }

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: googleapis/gcloud-mcp Stars: 796 Language: TypeScript Topics: Description: gcloud MCP server README (first 3000 chars): [](https://github.com/googleapis/gcloud-mcp/actions/workflows/presubmit.yml) [](https://github.com/googleapis/gcloud-mcp/blob/main/LICENSE) # gcloud MCP Server ☁️ The gcloud [Model Context Protocol (MCP)](https://modelcontextprotocol.io/docs/getting-started/intro) server enables AI assistants to easily interact with the Google Cloud environment using the gcloud CLI. With the gcloud MCP server you can: - **Interact with Google Cloud using natural language.** Describe the outcome you want instead of memorizing complex command syntax, flags, and arguments. - **Automate and simplify complex workflows.** Chain multiple cloud operations into a single, repeatable command to reduce manual effort and the chance of error. - **Lower the barrier to entry for cloud management.** Empower team members who are less familiar with gcloud to perform powerful actions confidently and safely. ## 📡 Available MCP Servers This repository also hosts other MCP servers in addition to the gcloud MCP server. An up to date list is below, and links to other Google Cloud MCP servers hosted outside of this repo are [here](#-other-google-cloud-mcp-servers). | MCP Server Name | Description | Package Name | Version | | :-------------- | :---------------------------------------------------------------------------- | :------------------ | :---------------------------------------------------------------------------------------------------------------------------------------- | | gcloud | Interact with Google Cloud via the gcloud CLI using natural language prompts. | `gcloud-mcp` | [](https://www.npmjs.com/package/@google-cloud/gcloud-mcp) | | observability | Access Google Cloud Observability APIs to query logs, metrics, and traces. | `observability-mcp` | [](https://www.npmjs.com/package/@google-cloud/observability-mcp) | | storage | Interact with Google Cloud Storage for bucket and object management. | `storage-mcp` | [](https://www.npmjs.com/package/@google-cloud/storage-mcp) | | backupdr | Interact with Google Cloud Backup and Disaster Recovery. | `backupdr-mcp` | [](https://www.npmjs.com/package/@google-cloud/backupdr-mcp) | ## 🚀 Getting Started ### Prere [ASSISTANT] { "what_it_is": "Official Google MCP server that lets AI assistants drive the gcloud CLI with natural language, plus sibling MCP servers for Observability, Cloud Storage, and Backup/DR.", "stack": ["TypeScript", "Node.js", "MCP", "gcloud CLI", "Google Cloud"], "why_mike_cares": "Low overlap — Mike's stack is Vercel/Netlify/Cloudflare/Supabase, not GCP, so a gcloud MCP doesn't plug into his agent fleet.", "verdict": "skip", "verdict_reason": "Mike doesn't run workloads on Google Cloud; the MCP pattern is already familiar and this one is locked to a platform he doesn't use." }

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: firecrawl/open-scouts Stars: 1282 Language: TypeScript Topics: ai-agents, alerts, automation, email-notifications, firecrawl, nextjs, open-source, openai, posthog, react, resend, supabase, tailwindcss, typescript, vercel, web-monitoring, web-scraping Description: 🔥 AI-powered web monitoring platform. Create automated scouts that search the web and send email alerts when they find what you're looking for. README (first 3000 chars): # Open Scouts Create AI scouts that continuously search the web and notify you when they find what you're looking for.  ## About Open Scouts is an AI-powered monitoring platform that lets you create "scouts" - automated tasks that run on a schedule to continuously search for and track information. Whether you're looking for new restaurants near you, monitoring AI news, or tracking any other updates, scouts work 24/7 to find what you need and notify you when they discover it. ## Tech Stack - **Next.js 15** - **React 19** - **TypeScript** - **Tailwind CSS v4** - **Supabase** (Database + Auth + Edge Functions) - **pgvector** (Vector embeddings for semantic search) - **Firecrawl SDK** (@mendable/firecrawl-js) - **OpenAI API** (AI Agent + Embeddings) - **Resend** (Email Notifications) ## Getting Started ### Prerequisites - Node.js 18+ - bun (default), npm, or pnpm - Supabase account ([supabase.com](https://supabase.com)) - OpenAI API key ([platform.openai.com](https://platform.openai.com)) - Firecrawl API key ([firecrawl.dev](https://firecrawl.dev)) - Resend API key ([resend.com](https://resend.com)) - for email notifications - Google Cloud Console account (for Google OAuth - optional) ### 1. Clone and Install ```bash git clone https://github.com/firecrawl/open-scouts cd open-scouts bun install # or: npm install / pnpm install ``` ### 2. Create Supabase Project 1. Go to [supabase.com](https://supabase.com/dashboard) 2. Create a new project 3. Wait for the project to finish provisioning ### 3. Enable Required Extensions In your Supabase Dashboard: 1. Go to **Database → Extensions** 2. Search for and enable: - `pg_cron` (for scheduled jobs) - `pg_net` (for HTTP requests from database) - `vector` (for AI-powered semantic search on execution summaries) - `supabase_vault` (for secure credential storage - usually enabled by default) ### 4. Set Up Environment Variables Create a `.env` file in the root directory by copying the example file: ```bash cp .env.example .env ``` Then fill in your actual values in the `.env` file. **The `.env.example` file contains all required environment variables with detailed instructions and direct links for where to obtain each API key.** ### 5. Run Database Setup First, link your Supabase project (required for syncing secrets): ```bash bunx supabase login # Login to Supabase CLI (one-time) bunx supabase link --project-ref <your-project-ref> # Find ref in Supabase Dashboard URL ``` Then run the setup script: ```bash b [ASSISTANT] {"what_it_is":"Open-source AI web monitoring platform that creates scheduled scouts to search the web via Firecrawl and send email alerts when matches are found.","stack":["Next.js 15","React 19","TypeScript","Tailwind CSS v4","Supabase","pgvector","Firecrawl SDK","OpenAI API","Resend","pg_cron"],"why_mike_cares":"Direct overlap: Firecrawl-first scraping rule, Next.js + Supabase + Vercel stack, scheduled agent patterns, and competitor/lead monitoring use cases for SEO and local service clients.","verdict":"check-it-out","verdict_reason":"Reference implementation from Firecrawl itself combining Mike's exact stack with scheduled scouts, pgvector summaries, and Resend alerts, ready to fork for competitor watch and lead-gen monitoring."}

[USER] You are an expert technical analyst for Mike Merlino, an AI agency operator and builder. Your job is to evaluate GitHub repositories and extract structured signal. Mike's domains: AI agents, LLM tooling, Claude Code, MCP servers, prompt engineering, SEO automation, GoHighLevel, Next.js, ShadCN, Vercel, Supabase, voice AI, agentic coding, Discord/Telegram bots, scheduler/cron patterns, Python automation, TypeScript. Return ONLY valid JSON: { "what_it_is": "one sentence plain English description", "stack": ["Technology1", "Technology2"], "why_mike_cares": "one sentence on overlap with Mike's work, or 'Low overlap'", "verdict": "check-it-out" | "worth-knowing" | "skip", "verdict_reason": "one sentence" } Verdict: - "check-it-out": direct overlap with Mike's active projects or tools he uses - "worth-knowing": interesting adjacent tool, may matter later - "skip": no clear overlap Be OPINIONATED. Not neutral. Repo: apache/ofbiz-framework Stars: 1039 Language: Java Topics: content, database, geospatial, groovy, hacktoberfest, https, java, javascript, network-server, ofbiz, plugins, web-framework, xml Description: Apache OFBiz is an open source product for the automation of enterprise processes. It includes framework components and business applications for ERP, CRM, E-Business/E-Commerce, Supply Chain Management and Manufacturing Resource Planning. OFBiz provides a foundation and starting point for reliable, secure and scalable enterprise solutions. README (first 3000 chars): <!-- Licensed to the Apache Software Foundation (ASF) under one or more contributor license agreements. See the NOTICE file distributed with this work for additional information regarding copyright ownership. The ASF licenses this file to you under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. --> # Apache OFBiz® [](http://www.apache.org/licenses/LICENSE-2.0) [](https://github.com/apache/ofbiz-framework) [](https://github.com/apache/ofbiz-framework/actions/workflows/gradle.yml) [](https://github.com/apache/ofbiz-framework/actions/workflows/codeql-analysis.yml)<!-- For the CodeQL badge to not get stuck on failing, you need to delete the last failing CodeQL workflow (AKA action)--> [](https://ci2.apache.org/#/builders?tags=%2BofbizTrunkFrameworkPlugins)<!--For the BuildBot badge to not get stuck on failure, ALL builds of all branches handled by https://svn.apache.org/repos/infra/infrastructure/buildbot2/projects/ofbiz.py need to pass--> [](https://qpkb254zxeu.montastic.io) [](https://github.com/apache/ofbiz-framework/actions/workflows/scorecard.yml) [](https://securityscorecards.dev/viewer/?uri=github.com/apache/ofbiz-framework) [](https://www.bestpractices.dev/projects/8708) --- Welcome to **Apache OFBiz**! A powerful [ASSISTANT] {"what_it_is":"Apache OFBiz is a Java-based open source enterprise resource planning framework with prebuilt ERP, CRM, e-commerce, supply chain, and manufacturing modules.","stack":["Java","Groovy","XML","JavaScript","Gradle"],"why_mike_cares":"Low overlap — heavyweight legacy Java ERP framework, opposite of Mike's Next.js/Supabase/AI agent stack and local service business focus.","verdict":"skip","verdict_reason":"Enterprise Java ERP with zero touchpoint to AI agents, LLM tooling, SEO automation, or Mike's TypeScript/Python ecosystem."}