rag-eval
01Skill Claude CodeCodex needs its repo
Run the retrieval regression gate against the current repo state and report whether a recent change helped, hurt, or held steady.
not rated 5 7d ago A 0 tokens
original MIT
ai-architect
02Agent Claude Code
Designs AI/agent systems (agent topology, prompt architecture, RAG design, eval gates, orchestration patterns, model tiering, memory/knowledge-graph design, autonomy guardrails). Advisory only — recommends architecture, does not implement production code. Use for agent design, prompt engineering, retrieval…
not rated 1 3d ago A 88 tokens
original MIT
architect
03Agent Claude Code
Audit and improve codebase architecture in one orchestrated workflow. Chains coupling analysis → orphan hunt → deepening opportunities → domain sharpening → critic gate → ADR recording. Read-only analysis; does not implement changes. Use when planning a refactor, evaluating architectural health, or audit-deep flagged…
not rated 1 3d ago A 62 tokens
original MIT
Agent Claude Code
Build ROI-ranked, deduped backlogs from parallel repo analysis to GitHub issues on a Project board. 8-phase composite: discover (parallel) → rank → propose (approval gate) → spec → plan → issues → board → snapshot. Use when starting a new work session on a repo or when "what should I work on" needs a structured answer.
not rated 1 3d ago A 0 tokens
original MIT
ci-fixer
05Agent Claude Code
Diagnose and fix failing GitHub CI on your own PRs. Fetches Actions logs, queries repo CI history for prior patterns, summarizes failures, drafts a fix plan, and implements after explicit approval. Hard stop if PR belongs to someone else or has human reviewer comments. Use when PR checks fail and you need root-cause…
not rated 1 3d ago A 74 tokens
original MIT
Agent Claude Code
Cloudflare edge platform specialist for Workers, Pages, D1 (SQLite), KV, R2, Durable Objects, Hyperdrive, Wrangler, and OpenNext (Next-on-Cloudflare). Audits architecture, bundle size, secret/binding strategy, database migrations, storage quotas, and edge runtime constraints. Use for CF infra decisions, deployment…
not rated 1 3d ago A 112 tokens
original MIT
Agent Claude Code
Expert code review specialist with severity-rated feedback, logic defect detection, SOLID principle checks, style, performance, and quality strategy.
not rated 1 3d ago A 29 tokens
original MIT
Agent Claude Code
Simplifies and refines code for clarity, consistency, and maintainability while preserving all functionality. Focuses on recently modified code unless instructed otherwise.
not rated 1 3d ago A 34 tokens
copy · 95% MIT
Agent Claude Code
Community strategy & engagement specialist for Discord and social channels. Designs announcements, moderation tone & policy, onboarding flows, event ideas, member retention strategies, and conflict de-escalation. Use for announcement drafting, community voice/moderation guidance, onboarding strategy, member-retention…
not rated 1 3d ago A 83 tokens
original MIT
critic
10Agent Claude Code
Work plan and code review expert — thorough, structured, multi-perspective (Opus).
not rated 1 3d ago A 20 tokens
copy · 92% MIT
debugger
11Agent Claude Code
Root-cause analysis, regression isolation, stack trace analysis, build/compilation error resolution.
not rated 1 3d ago A 21 tokens
copy · 88% MIT
Agent Claude Code
Artifact-only adversarial reviewer for DECISIONS and analysis (Opus). Reasons solely on the provided ARTIFACT + CONTRACT — has NO evidence-gathering tools by construction, so it cannot fabricate facts from logs/evals it ran itself. Use for research-and-decide Phase 2; use critic for code/plan reviews that need to…
not rated 1 3d ago A 80 tokens
original MIT
Agent Claude Code
Conduct bounded decision interviews using AskUserQuestion to resolve design, architecture, or product forks one at a time. Maps decision trees first, asks one fork per turn, detects 95% convergence, and produces a confirmed restatement. Use when a plan has knowable alternative paths and bounded choices will resolve…
not rated 1 3d ago A 70 tokens
original MIT
deep-auditor
14Agent Claude Code
Composite health audit that runs test-health, config-drift-detect, hook-effectiveness, security-audit, mcp-audit, plugin-audit, and socket-audit in parallel, reconciles into severity-ranked findings, cross-checks against prior decisions via RAG, and produces a prioritized remediation plan. Use for "is this project…
not rated 1 3d ago A 84 tokens
original MIT
designer
15Agent Claude Code
UI/UX Designer-Developer for stunning interfaces (Sonnet).
not rated 1 3d ago A 15 tokens
original MIT
Agent Claude Code
External Documentation & Reference Specialist.
not rated 1 3d ago A 9 tokens
copy · 91% MIT
Agent Claude Code
Analyze a proposed multi-agent workflow, Workflow() script, or active session for token waste and time bottlenecks. Catches model-tier mismatches, sequential work that should be parallel, re-read waste, and context bloat — scores each fix by impact with estimated savings. Use before spawning large agent fleets, when…
not rated 1 3d ago A 86 tokens
original MIT
explore
18Agent Claude Code
Codebase search specialist for finding files and code patterns.
not rated 1 3d ago A 13 tokens
copy · 92% MIT
Agent Claude Code
Web frontend motion craft specialist. Designs and implements UI animation that feels alive and welcoming without reading as "AI-made": orchestrated page-load, scroll-driven narratives, cursor/magnetic interactions, reveal/masking, 3D/WebGL depth, and micro-interactions. Owns timing/easing/spring tokens, native…
not rated 1 3d ago A 0 tokens
original MIT
git-master
20Agent Claude Code
Git expert for atomic commits, rebasing, and history management with style detection.
not rated 1 3d ago A 18 tokens
original MIT
Agent Claude Code
Capture active work state before budget runs low, switching projects, or ending a session. Writes a durable resume packet to /.claude/handoffs/ /latest.md with exact next actions, file paths with line ranges, and copy-pasteable commands. Use before context switches, approaching token budget, or end-of-day.
not rated 1 3d ago B 71 tokens
original MIT
Agent Claude Code
Move issues through a triage state machine — needs-triage → needs-info | ready-for-agent | ready-for-human | wontfix. Gathers context, reproduces bugs, grills for spec gaps, posts agent briefs or triage notes. Use when triaging a backlog of issues, evaluating a specific issue, or preparing issues for autonomous agent…
not rated 1 3d ago A 76 tokens
original MIT
marketing
23Agent Claude Code
Brand & content marketing strategist for positioning, messaging, campaign concepting, content strategy, and copywriting direction. Use for brand campaign strategy, content calendar planning, messaging frameworks, copy direction, funnel architecture, and positioning refinement — not for ad buying (paid-traffic), not…
not rated 1 3d ago A 74 tokens
original MIT
mcp-tool-dev
24Agent Claude Code
MCP development specialist. Expert in building Model Context Protocol servers and tools (registration, Zod schemas, transports) and in operating MCP gateways (routing, auth, rate limiting, health). Use when creating or debugging MCP tools/servers, or when working on gateway routing, authentication, and service…
not rated 1 3d ago A 66 tokens
original MIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: