THE 20 THE 20 THE 20
A curated flagship set of skills. Internally consistent, technically plausible, and well-structured — the "nobody else has" framing backed by implementation evidence.
Production-grade pipeline Each skill reads like an implementable module — agents, pipelines, guardrails, rendering, PRD conversion. Backed by real deployment logs, repo links, and battle-tested patterns.
Browse Type to filter • click chips to scope
ALL META AGENTS DESIGN PATTERNS COMMAND
01. Claudeception META PLAUSIBLE
Self-extracting learning system that creates new skills from sessions.
02. SkillForge META PLAUSIBLE
11-thinking-model skill router with regression questioning & evolution lens.
03. continuous_learning META PLAUSIBLE
Auto-detects learning opportunities from every interaction.
04. skill-scout META PLAUSIBLE
Hourly daemon that scans registries for new skills to absorb.
05. algorithm-to-prd-growth-strategy AGENTS VERIFIED
Reverse-engineer Twitter/TikTok/etc algorithms into growth PRDs.
06. create-comprehensive-prds-from-research AGENTS VERIFIED
Academic papers → implementation-ready PRD suites.
07. prd-to-ralph-conversion AGENTS VERIFIED
PRDs → autonomous agent executable JSON.
08. ralph-story-sizing AGENTS VERIFIED
Prevents context exhaustion in autonomous dev agents.
09. glassmorphism-anime-cyberpunk-design-system DESIGN VERIFIED
Full frosted glass + neon aesthetic system.
10. midjourney-style-iterative-generation-ux DESIGN VERIFIED
Generate 4 → pick → upscale workflow pattern.
11. remotion-terminal-promo-video DESIGN VERIFIED
Terminal-aesthetic promo videos with 3D transforms.
12. flux-lora-comprehensive-prompt-engineering DESIGN PLAUSIBLE
FLUX/SD LoRA training caption methodology.
13. telegram-bot-image-ui-rendering PATTERNS VERIFIED
Render rich UIs as PNGs for Telegram (PIL/Pillow).
14. claude-api-empty-text-response-debugging PATTERNS VERIFIED
Fix Claude API agentic loops returning nothing.
15. natural-language-tool-integration-pattern PATTERNS VERIFIED
Conversational AI ↔ tool control bridge.
16. python-bot-external-drive-io-bottleneck PATTERNS VERIFIED
Diagnose bot hangs from external drive I/O.
17. untracked-files-deployment-audit PATTERNS VERIFIED
Catch missing files before deploy crashes.
18. mandatory-testing-before-presenting COMMANDMENTS VERIFIED
Forces end-to-end testing before claiming "done".
19. no-placeholder-no-mock-data COMMANDMENTS VERIFIED
Bans stock content; enforces real data.
20. last30days-skill COMMANDMENTS PLAUSIBLE
Research last 30 days across Reddit/X/Web; output ready-to-paste prompts.
21. fastapi-starlette-version-mismatch PATTERNS VERIFIED
Fix ValueError in FastAPI middleware stack from Starlette incompatibility.
22. nextjs-tailwind-missing-globals-css PATTERNS VERIFIED
Fix unstyled Next.js pages when Tailwind classes exist but don't apply.
23. nextjs-postcss-commonjs-error PATTERNS VERIFIED
Fix hasOwnProperty / export default errors in postcss.config.js.
24. nextjs-14-config-typescript-limitation PATTERNS VERIFIED
Fix "Configuring via next.config.ts is not supported" in Next.js 14.
25. nextjs-inmemory-to-mongodb-migration PATTERNS VERIFIED
Migrate in-memory APIs to MongoDB preserving API contracts.
26. telegram-webhook-async-long-tasks PATTERNS VERIFIED
Handle 25s+ tasks in Telegram webhooks without timeout.
27. telegram-bot-conflict-multiple-instances PATTERNS VERIFIED
Resolve getUpdates conflicts from parallel bot instances.
28. telegram-bot-feature-porting PATTERNS VERIFIED
Port complex features between python-telegram-bot codebases.
29. telegraf-chat-username-type-guard PATTERNS VERIFIED
Fix TypeScript "username" error on Telegraf Chat type.
30. silent-async-error-antipattern PATTERNS VERIFIED
Detect webhooks that .catch() but never notify users.
31. async-job-polling-generic-pattern PATTERNS VERIFIED
Generic polling utility across Replicate/OpenAI/Google providers.
32. fstring-brace-escape-in-prompts PATTERNS VERIFIED
Fix NameError in f-string system prompts with JSON examples.
33. replicate-private-model-version-hash PATTERNS VERIFIED
Fix 404s when calling private Replicate models via API.
34. api-error-handling-user-facing-messages PATTERNS VERIFIED
Return detailed error context instead of generic "API failed".
35. railway-bot-not-responding-api-key PATTERNS VERIFIED
Debug bots that deploy but don't respond on Railway.
Convert PRDs to prd.json for the Ralph autonomous agent system.
Generate Product Requirements Documents for new features.
38. create-comprehensive-prds-for-ralph AGENTS VERIFIED
Consolidate scattered work into actionable Ralph tasks.
39. react-three-fiber-dashboard DESIGN VERIFIED
Build 3D dashboards with R3F, Framer Motion, and Blender GLBs.
40. structured-data-to-visual-ui-telegram DESIGN VERIFIED
Auto-render plans/roadmaps as visual PNGs for Telegram delivery.
41. telegram-bot-image-ui-rendering DESIGN VERIFIED
PIL-based pixel-perfect dashboard rendering for Telegram bots.
42. nextjs-markdown-documentation-system-pattern DESIGN VERIFIED
Build beautiful docs sites that render markdown with custom styling.
43. git-push-large-files-history-cleanup PATTERNS VERIFIED
Fix "GH001: Large files detected" by cleaning git history.
44. git-unrelated-histories-resolution PATTERNS VERIFIED
Resolve "refusing to merge unrelated histories" errors.
45. git-push-repository-not-found-remote-mismatch PATTERNS VERIFIED
Fix push failures when SSH works but remote doesn't match.
46. humanizer COMMANDMENTS VERIFIED
Remove signs of AI-generated writing. Based on Wikipedia's AI writing guide.
47. multi-repo-todo-audit META VERIFIED
Audit multiple repos for TODOs, uncommitted changes, and tech debt.
48. attn-money-commandments-integration AGENTS VERIFIED
Connect tactical strategies to the ATTN.MONEY commandments system.
Showing 48 of 48 skills • Each skill has a SKILL.md source file with full implementation details.
Verification rubric how to "prove" each skill
This registry reads like an operator dashboard. Evidence attached per skill.
Proof package (per skill) EASY repeatable
1) One-sentence spec • 2) CLI/API example • 3) screenshot/log • 4) metric (time saved, success rate) • 5) link to source.
Battle-tested positioning EARNED Every skill tagged VERIFIED has been deployed in production, debugged under pressure, or extracted from a real shipping codebase.
Score fields optional maturity (0–100), impact (hrs saved/wk), reliability (success %), coverage (features), risk (failure modes).
Community & token SOON Memecoin / reward token for early skill users and contributors is in the works. No wallet required to use the registry — token is optional community layer.
Drop strategy (short)
Don’t gate the registry — token is optional. Solana (pump.fun) or Base; airdrop to early visitors / contributors. One claim page; link from here. No “connect wallet” on main site. Message: reward for community, not investment.
Visual language: mono type, tight hierarchy, glass panels, neon accents. Terminal-grade clarity for operator-level tools.