AGENTS.md · sickn33/agentic-awesome-skills
AGENTS.md from sickn33/agentic-awesome-skills: 109 lines covering Project Structure & Module Organization, Build, Test, and Development Commands, Coding Style…
Template by sickn33
Chalk Labs
Need help deploying AGENTS.md · sickn33/agentic-awesome-skills?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode, Claude Code. Read the repository before you give any tool access to your data.
curl -fsSL https://raw.githubusercontent.com/sickn33/agentic-awesome-skills/main/AGENTS.md -o AGENTS.md
Read it and adapt it to your project before use.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: prompt-engineering
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
What's in it
An AGENTS.md file: instructions coding agents (Codex, Cursor, Copilot, Gemini CLI and others) read before working in a repo.
- For
- Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode, Claude Code
- Length
- 109 lines
- File
AGENTS.md
Sections
- Project Structure & Module Organization
- Build, Test, and Development Commands
- Coding Style & Naming Conventions
- Testing Guidelines
- Commit & Pull Request Guidelines
- Agent-Specific Instructions
- Learned User Preferences
- Learned Workspace Facts
- Mandatory release cycle barrier
First lines (shown under the repo's MIT license)
# Repository Guidelines ## Project Structure & Module Organization This repository publishes an installable library of agent skills and plugin bundles. Canonical skill sources live in `skills/<skill-id>/SKILL.md`; use lowercase, hyphenated skill IDs. Mirrored plugin distributions live under `plugins/`. Contributor and user docs live in `docs/`; localized docs live in `docs_zh-CN/` and `docs/vietnamese/`. Maintenance scripts and tests are in `tools/scripts/` and `tools/scripts/tests/`. The hosted catalog app is in `apps/web-app/`. Registry outputs such as `CATALOG.md`, `skills_index.json`, and `data/catalog.json` are generated artifacts. Editorial inputs in `data/` remain source files; use `tools/scripts/generated_files.js` for the exact generated-file classification. ## Build, Test, and Development Commands - `npm ci`: install root dependencies for scripts and validation. - `npm run validate`: validate skill frontmatter, required sections, and schema rules. - `npm run security:docs`: run safety checks for command, install, credential, and network guidance. - `npm run test`: run the repository script test suite. - `npm run build`: regenerate core indexes and build the catalog data. - `npm run app:install`: install `apps/web-app` dependencies. - `npm run app:dev`: start the local Vite catalog app. - `npm run app:build`: build and prerender the catalog app. For ordinary external contributions, fetch the official upstream and run `npm run pr:preflight -- --contributor --base upstream/main --head HEAD`. For maintainer topic PRs, run `npm run pr:preflight -- --base origin/main --head HEAD` from the clean committed source head. It runs the component policy in an isolated checkout; unknown and policy changes require the full suite. ## Coding Style & Naming Conventions Use Markdown for skills and docs, JavaScript/Node for most tooling, and Python for audits and sync helpers. Keep skill directories lowercase with hyphens, for example `skills/my-awesome-skill/SKILL.md`. Start new skills from `docs/contributors/skill-template.md`; include frontmatter, `## When to Use`, examples, and limitations. Keep generated-file edits out of community PRs unless doing maintainer release or sync work. ## Testing Guidelines Tests live mainly in `tools/scripts/tests/` and use Node assertions or Python `unittest`. Name new tests after the behavior under test, for example `installer_filters.test.js` or `test_validate_skills_strict.py`. Run targeted tests during development, then run the relevant npm scripts above. Web app changes should also run `npm run app:test` or `npm run app:test:coverage`. ## Commit & Pull Request Guidelines History uses conventional-style subjects such as `feat: add ...`, `fix: refresh ...`, `docs: add ...`, and `chore: release ...`. Keep commits focused. PRs must use the default template, include the Quality Bar Checklist, link an issue when applicable, and allow maintainer edits. Source PRs should avoid generated registry artifacts; CI enforces this source-only contract. ## Agent-Specific Instructions Respect deeper `AGENTS.md` files inside skill subtrees. When changing canonical skill content that is mirrored under `plugins/agentic-awesome-skills/` or `plugins/agentic-awesome-skills-claude/`, check whether mirrors must be synchronized. For release work, follow the scripted `release:prepare` and `release:publish` flow rather than hand-editing version surfaces. ### Current-Base Instruction Guard Repository instructions must match the exact Git base used for the task. After creating a clean clone, worktree, or topic branch, re-read that base's `AGENTS.md`, `.github/MAINTENANCE.md`, canonical maintainer skill, and `package.json`; those files supersede instructions inherited from the checkout that launched the task. Every command, script, reviewer, or gate described as mandatory must exist on the current task base. If it is absent, do not recover or execute it from another branch, worktree, stash, installed copy, or historical commit. Treat the mismatch as evidence that the procedure may have been retired, inspect `origin/main` and the relevant removal history, then follow the current-base contract or report the unresolved conflict. ### Mandatory Maintainer Workflow For every repository maintenance sweep, PR merge batch, maintainer-side PR repair, canonical synchronization, combined Security/Quality cleanup and merge, or tag/release request, **always invoke and follow the `antigravity-maintainer-batch-release` skill before triage or mutation**. If the client has not installed or discovered that skill, read and follow the repository-canonical copy at `skills/antigravity-maintainer-batch-release/SKILL.md`. This is a hard gate, including when the user asks for direct merges or a direct update to `main`; do not substitute a generic Git or GitHub workflow. Treat `main` as pull-request-only; use `npm run merge:batch` with strict protection and no admin bypass. Follow [the maintenance contract](docs/maintainers/maintenance-contract.md) for resumable review, release-only snapshots, immutable artifact publication and independent catalog revisions. Ordinary merges generate ephemeral validation data; general tracked distribution refresh belongs to protected release PRs. The maintenance contract defines a narrow owner-only exception for byte-identical canonical lockfile replicas during an explicitly requested no-release dependency-security repair; all checks and strict protection remain required. Every stable or prerelease version must finish with the full-release-alignment gate. Verify the immutable released commit and public bytes even if main advances. Report personal MCP host updates and actual session reload separately. Only authorized existing host entries may be changed. #### Skill Content Review Gate For every canonical `SKILL.md` change or tracked bundle-file change, run `npm run validate`, `npm run validate:references`, `npm run security:docs`, and the relevant tests. Inspect semantics, safety, provenance, declared risk, limitations, and every bundled file. The official merge gate remains a truthful Tessl `review` or a maintainer review attested to the exact full head SHA; heuristic local scores and inferred risk labels are not merge authority. Reviewed fork bundle exceptions are restricted to the protected-base ledger in `tools/config/reviewed-fork-skills.json` and the current maintainer skill. They bind a complete previously reviewed skill tree and still require exact-current-head attestation, all required checks and strict protection; no general script allowlist or PR-controlled ledger is authorized. ## Learned User Preferences - For maintainer sweeps, PR merges, issue closure, and releases, follow the canonical `antigravity-maintainer-batch-release` skill together with `.github/MAINTENANCE.md`; do not substitute a generic Git or GitHub workflow. - When executing an attached plan, implement the plan without editing the plan file; use existing todos instead of recreating them. - Optional TypeSafe Jev tooling must accelerate triage and quality; never add mandatory steps to `merge:batch`, CI, or branch protection. - Keep Jev usage within a small monthly TypeSafe budget by capping skills per run and treating output as advisory only. - Store TypeSafe API keys only in gitignored `.env.local`; rotate any key exposed in chat, logs, or commits. - Before committing new maintainer tooling, run a full validation round on a real open skill PR (worktree, deterministic checks, and Jev smoke when applicable). - Release changelog and GitHub release notes must match the actual tag diff (skill counts, catalog totals, contributor thanks); emphasize catalog skills in user-facing notes, not maintainer-only tooling. - Respond in Italian when the user writes maintainer or release requests in Italian. ## Learned Workspace Facts - `npm run maintainer:jev-hints` evaluates skill content from the `--head` git ref via `git show`; pass `--repo` with a PR worktree when changed skills are not present on the current checkout. - Jev hints default to five skills per run, exit 0 with a skip message when `TYPESAFE_API_KEY` is unset, and never satisfy Tessl or `--reviewed-head` skill review. - Maintainer documentation for Jev lives in `docs/maintainers/jev-hints.md`; the upstream TypeSafe agent skill is installed under `.agents/skills/typesafe-ai/` via `npx skills add typesafe-ai/skills --skill typesafe-ai`. ### Maintainer state and attribution Use private external review receipts and batch journals as described in the maintenance contract. They never replace current protection or exact-head evidence. Update code, rules, canonical skill, installed skill and negative regressions together when the user requests a workflow change. Use `maintenance:cycle start/intake/merge/verify` for a measured source batch (its merge delegates to `merge:batch`); finish this no-release cycle before any authorized release preparation; read `docs/maintainers/maintenance-cycle.md`. Record blockers once, continue independent PRs and reuse the existing host verifier. Do not run release steps when no release was requested. ## Mandatory release cycle barrier A release must never proceed with an unresolved PR in the recorded batch. Every batch PR must be merged into protected main or closed without merge with an owner rationale posted **before** closure and bound to its exact head. A policy blocker is not completion and never authorizes bypassing merge protection or closing an otherwise valid contribution merely to unlock publication. Use `npm run maintenance:cycle -- close --state-dir /absolute/private/state --prs 123 --expected-head FULL_40_CHARACTER_SHA --reason 'Substantive reason, including the replacement PR when applicable'`. This posts the required comment before closing; uncertain writes stop without a retry. Plain undocumented closure does not satisfy verification. Merged PRs use the existing protected merge path. Run `maintenance:cycle verify` successfully on exact current main before `npm run release:prepare -- X.Y.Z --state-dir /absolute/private/state`. Preparation fails before branch creation or version changes without completed cycle evidence. The release PR carries `data/release-cycle.json`; publication and the npm workflow independently re-read live PR states, closure comments, protected-main ancestry and successful CI/CodeQL for the verified source commit. Reopened PRs, absent proof or failed checks block publication. Reverify if main moves before preparation. Never shrink the recorded scope to hide blockers. For a genuinely release-only task with no source PR batch, explicitly start a cycle with `--prs none` and verify it; no implicit no-cycle bypass is supported. The original batch inventory must include every PR in the authorized scope. For contribution readiness, follow `docs/maintainers/contribution-intake.md`. Ask authors to resolve substantive reproducible submission defects and provide truthful scoped evidence; retain complete-tree exact-head semantic review. Restricted bundles need a reviewed maintainer path, not relaxed fork policy. Canonical skill-tree PRs must pass mandatory evidence admission in `pr-policy` and current-body revalidation in `merge:batch`. Body/draft drift stops an active merge; content attestations cannot waive incomplete evidence. Keep author requirements proportionate and free of mandatory paid tools or unsafe live tests. For release process repairs, validate the full shared candidate path through `npm run release:rehearse -- X.Y.Z --state-dir /absolute/private/completed-cycle`. Keep changes in an isolated topic checkout while another release is active; do not modify its source, state or installed skill. Rehearsal never grants publication authority.
Signals
- Stars
- 48,428
- Forks
- 7,039
- Open issues
- 0
- Last commit (this folder)
- Oct 24, 2026 (7 days ago)
- Last release
- Oct 24, 2026 (6 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from sickn33
- 0.665score
agentphone
sickn33
Build AI phone agents with AgentPhone API.
SkillAFreshness: A★ 48.4KClaude Code +9 - 0.665score
ai-seo
sickn33
Optimize content for AI search and LLM citations across AI Overviews, ChatGPT, Perplexity, Claude, Gemini, and similar systems.
SkillAFreshness: A★ 48.4KClaude Code +9 - 0.665score
agent-evaluation
sickn33
Evaluate agent behavior with versioned cases and explicit verifiers.
SkillAFreshness: A★ 48.4KClaude Code +9 - 0.665score
agent-memory-systems
sickn33
Design tenant-scoped memory, retrieval and retention for agents.
SkillAFreshness: A★ 48.4KClaude Code +9 - 0.665score
add-app-clip
sickn33
Add an iOS App Clip target to an Expo app.
SkillAFreshness: A★ 48.4KClaude Code +9 - 0.665score
angular-best-practices
sickn33
Angular performance optimization and best practices guide.
SkillAFreshness: A★ 48.4KClaude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get AGENTS.md · sickn33/agentic-awesome-skills running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
AGENTS.md · sickn33/agentic-awesome-skills. Agent Mart. https://agentmart.dev/templates/sickn33-agentic-awesome-skills-agents-md (last updated 2026-10-24)