Claude Skills
The open format is called Agent Skills and works in Claude Code, Codex, Cursor and other agents — most people know it as Claude Skills.
Every Agent Skill we could find on GitHub, deduplicated by content. 80 149 files from 1 774 authors, of which 62 489 are unique — the rest is the same skill repackaged into someone else's repository. For each one: what it weighs in tokens, whether it ships runnable scripts, and which MCP servers it needs.
Finish an already-merged branch. Proves the branch is an ancestor of the fetched default, classifies leftover artifacts as unique / redundant / superseded, returns to a clean --ff-only default checkout, and deletes the merged local branch — every discard confirmed per item. Routed to by /ca-cleanup.
The behavior-preserving restructure gate. Routed to by /refactor for a rename, extract, inline, move, dedup, or internal-implementation swap. Six gated phases prove behavioral parity through unmodified pre-existing tests; any diff that classifies as `feat` is not a refactor and is routed to tdd. A modified pre-existing test is rejected as evidence — it is a behavior change in disguise.
The secret-source gate. Routed to when changed code reads, writes, or passes a secret — API key, token, password, connection string, signing key, certificate, or any value that grants access. Validates that every secret comes from the approved store and never lands in source, log, test fixture, error response, image, or LLM prompt. The auth-crypto-reviewer agent is dispatched as the reviewer.
Optional, opt-in STRIDE threat pass for a sensitive feature — invoked deliberately via /threat-model, never forced on ordinary changes. Walks the change's attack surface and security boundaries (governed by <project-root>/.codearbiter/security-controls.md), surfaces threats and unmitigated gaps, and MAY dispatch security-reviewer or auth-crypto-reviewer. Not a routine gate; it can hard-STOP only on a genuinely critical unmitigated threat it surfaces.
The single permitted path to a version tag, for any target declared in the project's .codearbiter/release-targets.md. Routed to when the user invokes /release on a non-default branch with a green suite. Takes the declared target as its one argument, derives the SemVer bump from Conventional-Commits history since that target's last tag, rolls the commits into that target's CHANGELOG, writes an annotated tag in that target's namespace, and on authorization publishes it as a GitHub Release with the changelog section as its notes. A release commit, if needed, routes through commit-gate; the tag and Release are never published without explicit authorization.
The authoring gate for new skills. Routed to when the user invokes /new-skill "<gap>". Five gated phases — gap evidence, scope, authoring, self-review against the v2 house style, routing integration. A new skill is not written until an existing one is proven not to cover the gap, and not shipped until it carries gated phases, hard rules, and a routing entry. Every authored skill matches the v2 format (frontmatter name+description, # name, Pre-flight, Phase N · gate, Hard rules).
The implementation engine. Routed to by /sprint (full plan, autonomous) and by executing-plans (scoped batch, checkpoint-gated). One fresh subagent per task — test-first via tdd — followed by spec-compliance review, quality review, and fresh-run verification. No single context accumulates drift, and nothing is accepted on a subagent's word.
The test-first gate. Routed to by /feature (after the spec is approved), /fix, and /refactor before any implementation code is written. Six gated phases — obligation scan, red, green, obligation verify, coverage, lint. No feature code exists before Phase 1 clears; nothing reaches commit-gate until all six are green.
The deep, rarely-convened whole-codebase audit lane. Routed to when the user invokes /ca-tribunal. Seven gated phases — cost/model, map, roster dispatch, triage, report, approval+filing, telemetry. Costs on the order of millions of tokens; proceeds only after the user acknowledges the estimate; never a required gate; nothing filed or sent without explicit authorization.
The spec-to-plan bridge. Routed to by /feature once the brainstormed spec is approved, and by /sprint before execution. Decomposes the spec into 2–5 minute tasks, each carrying its exact file path(s) and a concrete verification step that maps to a tdd obligation. Writes the plan to .codearbiter/plans/<slug>.md, ordered with dependencies flagged and an MVP slice identifiable. Nothing executes until every task has a path and a verification and the task set covers every acceptance criterion.
Report the health of Architecture Decision Records — aged, unchallenged, supersession candidates, unresolved CONFIRM-NN. Read-only.
Author a numbered, dated, user-attributed Architecture Decision Record under .codearbiter/decisions/.
Exit maintainer dev mode — restore orchestration, remove the dev marker, log the exit.
Vet a new or changed third-party dependency for license, provenance, and supply-chain risk before any install runs.
Assemble the governance record for a range — commits, overrides, ADRs, sprint auto-decisions, open questions, checkpoint findings — into one dated audit packet. Read-only.
Lightweight Q&A about the project — answer from context and return, no routing, no state change.
Sanctioned lane for non-behavioral work — docs-only edits, dependency bumps, reverts. Type-scaled gates; no TDD demanded of prose.
Periodic multi-reviewer sweep of the whole codebase — surfaces a triaged checkpoint report.
Finish an already-merged branch — classify the leftover artifacts, return to a fast-forwarded default checkout, and delete the merged local branch. Every discard confirmed per item; ancestry proven, never assumed.
Show the codeArbiter command catalog — the public command list and what each routes to.
Run the full commit gate — the only sanctioned path to a git commit.
Stop everything and surface a rule conflict — persona vs. docs vs. code. Present both sides and the conflict-hierarchy level; the user resolves. No silent reconciliation.
Brownfield back-fill — scout an existing codebase and populate .codearbiter/, then lock it initialized.
Investigate-then-decide root-cause analysis for a defect whose cause is unknown (distinct from /fix, which assumes a known bug). Five gated phases: capture, hypothesize, gather, decide, hand off. Investigation only, no code changes; exits to /fix, /adr, or a no-action close.
Maintainer override — suspend orchestration to edit codeArbiter itself. Env-gated (CODEARBITER_DEV=1), entry/exit logged to overrides.log.
The greenfield decomposition interview. Routed to at startup when .codearbiter/CONTEXT.md lacks the <!--INITIALIZED--> body marker and no source code exists, or when the user invokes /decompose. A senior-architect persona drives a six-layer interview, persists every layer to disk so a context reset loses nothing, then populates .codearbiter/ and locks it initialized. No project-state doc is written before the layers are solid; orchestration does not resume until the lock is set.
Start a feature: brainstorm a spec, get it approved, then drive it test-first through the pipeline. The one entry to implementation.
Verify the active host install, package, command ownership, enforcement, wrapper self-test, and active-dispatch coverage gap. Read-only.
Fix a confirmed bug: a failing regression test first, then a minimal fix, then the rest of the tdd gates.
Opt this repo into codeArbiter — scaffold the root-level .codearbiter/ state store.
Read-only 3-metric governance glance — override rate, small-lane rate, sprint low-confidence ratio — each with a trend arrow vs. the prior 20-commit window.
Author a new codeArbiter skill: prove the gap is real, get the spec approved, then write it.
Open a pull request the only sanctioned way — clear every BLOCK-level review finding, then stage the PR. Never a direct write to the default branch.
Sanctioned, logged bypass of a gate or hard rule — one audit line, then proceed.
Zero-onboarding, read-only dry-run of the reviewer fleet against the current uncommitted diff. Predicts reviewers, runs the state-free secret scan, writes nothing.
Trim transcript clutter to extend session lifetime — analyze, prune a copy, or toggle the after-each-turn service. Dry-run by default; gains land at resume/compaction, not the current turn.
SMARTS arbitration — reconcile architectural artifacts against the scaffold and prior decisions; every variance resolved by an explicit, user-attributed choice.
The behavior-preserving restructure gate. Routed to by /refactor for a rename, extract, inline, move, dedup, or internal-implementation swap. Six gated phases prove behavioral parity through unmodified pre-existing tests; any diff that classifies as `feat` is not a refactor and is routed to tdd. A modified pre-existing test is rejected as evidence — it is a behavior change in disguise.
The single permitted path to a version tag, for any target declared in the project's .codearbiter/release-targets.md. Routed to when the user invokes /release on a non-default branch with a green suite. Takes the declared target as its one argument, derives the SemVer bump from Conventional-Commits history since that target's last tag, rolls the commits into that target's CHANGELOG, writes an annotated tag in that target's namespace, and on authorization publishes it as a GitHub Release with the changelog section as its notes. A release commit, if needed, routes through commit-gate; the tag and Release are never published without explicit authorization.
Autonomous sprint — one interactive spec gate, then plan-to-PR execution with every auto-decision SMARTS-scored and logged. Hard gates remain true stops.
Exploratory spike on a throwaway branch — answer a named question with disposable code. Never merges; exits to a findings note or /ca-feature.
Review a diff with the reviewer fleet, funneled to one triaged verdict. Targets the current working diff, a path, or an inbound GitHub PR.
Daily repo hygiene — review the day's repo state, then perform the cleanups under per-action confirmation. Fast-forward only, never destructive without a yes.
Show the project's current state at a glance — stage, open tasks, open questions, overrides since the last checkpoint, current branch. Read-only.
The sanctioned task-board mutator — add a queued task, start one (flips to in-progress and stamps the date, minting a dotted ID on pick-up), or mark an in-progress task done. The only blessed write to open-tasks.md.
Opt-in lightweight STRIDE pass for a sensitive feature before implementation. Not a routine gate — invoke it when a change warrants security thought.
The deep, rarely-convened whole-codebase audit lane. Routed to when the user invokes /ca:tribunal. Seven gated phases — cost/model, map, roster dispatch, triage, report, approval+filing, telemetry. Costs on the order of millions of tokens; proceeds only after the user acknowledges the estimate; never a required gate; nothing filed or sent without explicit authorization.
Run Claude Code INSIDE a ca-sandbox box (`--with-claude`). Routed to when the user wants an agent loop running against an isolated, ephemeral sandbox rather than the host. Authenticates via an env-injected CLAUDE_CODE_OAUTH_TOKEN with no host bind of ~/.claude; the image pins the CLI and disables the autoupdater; HOME is backed by a named volume so the .claude state persists across restart. Five gated phases — posture, image, token, run, teardown. The hard default is offline or Anthropic-domains-only egress, and the token volume is NEVER co-mounted with an untrusted-code run; both are enforced, not advised.
Watch a PR's CI to completion — diagnose on red, notify and offer the merge on green. Never auto-merges.
The lifecycle gate for a local Codespace-equivalent sandbox. Routed to when the user invokes /ca-sandbox:sandbox to pull an untrusted repo into an ephemeral, host-FS-isolated Docker container, or any of the interaction commands (/ca-sandbox:sandbox-shell, /ca-sandbox:sandbox-exec, /ca-sandbox:sandbox-cp, /ca-sandbox:sandbox-destroy) against an existing box. Five gated phases — pre-flight, clone+build, isolated run, interact, teardown. The load-bearing invariant is structural: untrusted code in the box can never reach the host filesystem (no bind mount, no docker socket, never --privileged, cap-drop ALL, non-root, read-only root). Network defaults to offline; egress out is host-initiated only. Every object is labeled ca.sandbox=1 and torn down on exit.
The only path to a commit. Routed to when the user invokes /commit or otherwise instructs codeArbiter to persist staged changes. Nine gated phases — permission, branch, classification, verification (test/lint/secrets), behavioral proof, diff review, selective stage, message, commit. Nothing reaches version control without clearing every gate; "it looks good" is not authorization.
The Socratic spec-refinement front of /feature, and the planning front of /sprint. Routed to BEFORE any code — it takes a one-line idea and drives it to an approved, concrete spec with testable acceptance criteria. Four gated phases — frame, refine, write, approve. No implementation and no handoff to tdd until the spec is on disk and approved; each acceptance criterion becomes one tdd Phase 1 obligation.
The brownfield back-fill. Routed to by /create-context, and by startup when .codearbiter/CONTEXT.md lacks the <!--INITIALIZED--> body marker but source code exists. Six gated phases — pre-flight, scout dispatch, synthesis, gap interview, write, lock. Reads the existing codebase through parallel scouts, drafts every surviving project-state doc, resolves gaps with the user, and locks the project as initialized.
Author and track Architecture Decision Records. Routed to when the user invokes /adr to record a new decision or /adr-status to list ADR health. Authors numbered, dated, user-attributed ADRs under .codearbiter/decisions/, maintains supersede chains, and reports status read-only. Never authors an ADR as its own judgment — every ADR carries explicit user attribution.
The banned-primitive gate. Routed to when changed code hashes, signs, encrypts, derives keys, generates security-relevant randomness, configures TLS, or imports a crypto library. Rejects broken primitives, disabled TLS verification, and home-rolled crypto; the approved-primitive list lives in security-controls.md. The auth-crypto-reviewer agent is dispatched as the reviewer.
Reconcile the project's architectural artifacts against the scaffold and prior decisions, then present each variance as a SMARTS analysis for the user to decide. Routed to when the user asks to arbitrate, reconcile, or consolidate architectural context, requests a variance report, mentions ADR conflicts, or asks which downstream artifacts the current state supports. Never decides alone — every arbitration is user-attributed and logged.
The checkpoint coordinator for /feature. Routed to by /feature once a writing-plans plan exists. Groups tasks into batches, delegates each batch to subagent-driven-development (fresh author agent per task, full review chain, fresh verification), then stops for a human checkpoint before the next batch. The checkpointed counterpart to /sprint's autonomous run.
The terminal step of /feature and /sprint. Routed to once commit-gate has cleared, to decide the branch's fate — merge via PR, open a PR, or discard. Direct merge to the default branch is forbidden; every change lands through a PR. Under /sprint the skill auto-selects "open PR" and surfaces the merge decision to the user.
Finish an already-merged branch. Proves the branch is an ancestor of the fetched default, classifies leftover artifacts as unique / redundant / superseded, returns to a clean --ff-only default checkout, and deletes the merged local branch — every discard confirmed per item. Routed to by /ca:cleanup.
The secret-source gate. Routed to when changed code reads, writes, or passes a secret — API key, token, password, connection string, signing key, certificate, or any value that grants access. Validates that every secret comes from the approved store and never lands in source, log, test fixture, error response, image, or LLM prompt. The auth-crypto-reviewer agent is dispatched as the reviewer.
Claude Skills — questions
Answers built from the skills we actually parsed.