settings — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited settings (Hook) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
<div align="center">
One `git checkout` drops a shared AI coding setup (skills, subagents, MCP servers, OpenSpec scaffolding) into any project.
agent-standards quickstart demo
</div>
For developers and teams running multiple AI coding agents across multiple repos. Every project drifts (different prompt files, skill versions, MCP configs, subagent definitions). This repo is the single source: import once, pull updates with one git fetch. Used across every project in the portfolio.
Consumer projects pull only production-ready content: skills, generated subagent files, the tooling doc, and config templates. The canonical subagents/ source and tools/ generator stay here.
The canonical source is .agents/skills/, the same directory GitHub Copilot and Cursor read natively for agent skills. The layout follows the cross-tool AGENTS.md and SKILL.md conventions, so the setup stays portable across the supported agents rather than tied to one tool.
git config core.symlinks truegit remote add agent-standards https://github.com/Lukk17/agent-standardsgit remote set-url --push agent-standards no_pushgit fetch agent-standardsgit checkout agent-standards/master -- .agents .claude .opencode .codex .kilocode .github/agents .github/copilot-instructions.md.example .vscode/mcp.json.example docs/AGENT_TOOLING.md docs/MCP_SETUP.md docs/AGENTS-UPDATE.md AGENTS.md.example opencode.json.example .mcp.json.examplecp AGENTS.md.example AGENTS.mdThat's it. Open the project in Claude Code, Kilo Code, OpenCode, Codex, or GitHub Copilot and skills are live.
Optional next step: copy .mcp.json.example to .mcp.json (Claude Code), opencode.json.example to opencode.json (OpenCode), .kilocode/mcp.json.example to .kilocode/mcp.json (Kilo Code VS Code extension), and .vscode/mcp.json.example to .vscode/mcp.json (GitHub Copilot in VS Code) to enable the shared MCP server set.
Human-side MCP setup lives in docs/MCP_SETUP.md.
Full walkthrough lives in docs/AGENT_TOOLING.md.
graph LR
subgraph AS[agent-standards repo]
ASK[.agents/skills/<br/>canonical SKILL.md files]
SU[subagents/<br/>canonical templates]
GEN[tools/gen_subagents.py]
SU --> GEN
end
AS -->|git checkout production-ready dirs only| P[Your project]
P --> SK[.agents/skills/]
P --> CA[.claude/agents/<br/>generated]
P --> OA[.opencode/agents/<br/>generated · shared by Kilo]
P --> AG[AGENTS.md]
SK -.symlink.-> CL[.claude/skills/]
SK -.symlink.-> OC[.opencode/skills/]
SK -.symlink.-> CX[.codex/skills/]
CL --> Claude[Claude Code]
SK --> Kilo[Kilo Code]
OC --> OpenCode[OpenCode]
CX --> Codex[Codex CLI]
CA --> Claude
OA --> Kilo
OA --> OpenCode
AG --> Claude
AG --> Kilo
AG --> OpenCode
AG --> Codex
P --> GH[.github/agents/<br/>generated]
GH --> Copilot[GitHub Copilot]
SK --> Copilot
AG --> Copilot
classDef src fill:#ede0ff,stroke:#7c3aed,color:#1a1033
classDef consumer fill:#fff7e0,stroke:#d97706,color:#1a1033
classDef tool fill:#e0f2fe,stroke:#0284c7,color:#1a1033
class AS,ASK,SU,GEN src
class P,SK,CA,OA,AG,CL,OC,CX,GH consumer
class Claude,Kilo,OpenCode,Codex,Copilot toolTwo canonical sources in this repo: .agents/skills/ and subagents/. Consumer projects receive only the production-ready dirs: .agents/skills/, generated .claude/agents/, generated .opencode/agents/ (which Kilo Code reads natively, so no separate Kilo directory), and generated .github/agents/ for GitHub Copilot. The subagents/ source and the tools/ generator never ship to consumers.
Full repository layout, agent compatibility matrix, and per-agent instruction precedence: docs/repository-layout.md.
SKILL.md files. Source of truth, shared by every agent, shippedto consumers. Browse the catalogue to see what's available.
gen_subagents.py and pyproject.toml. Emits.claude/agents/, .opencode/agents/, and .github/agents/.
AGENTS.md, a skills/ symlink, generated agents/.
openspec init creates the openspec/ directory in the consumerproject; commands land in each tool's native dir. See docs/AGENT_TOOLING.md for the full workflow.
Lukk17/openspec-schemas ships an e2e-runbooks schema for spec-driven end-to-end capability tests through OpenSpec's lifecycle (/opsx:new --schema e2e-runbooks). Install steps live in docs/AGENT_TOOLING.md.
AGENTS.md.example is the template you copy into a new project.
*.agent.md subagents, and.github/copilot-instructions.md.example is the template you copy to .github/copilot-instructions.md so local VS Code and IntelliJ chat pick up the standards.
.kilocode/mcp.json.example, .vscode/mcp.json.example, and docs/MCP_SETUP.md for the human setup.
Server list and configuration live in docs/MCP_SETUP.md. All entries use ${VAR} or {env:VAR} substitution with defaults baked in.
| Approach | What you get | Trade-off |
|---|---|---|
Copy-paste a single AGENTS.md | Fast start, one file | Drifts the moment upstream changes; no skills, no subagents, no MCP |
| Personal dotfiles repo | Total control | Manual sync; not designed for cross-tool agent setup |
| Tool-native rules (Cursor rules, Claude Projects) | Works in that tool | Locks you in; no portability across the supported agents |
| Awesome lists / prompt collections | Inspiration | Curated reading, not an installable setup |
| agent-standards | One git checkout brings skills, subagents, MCP servers across every supported agent | Adds an agent-standards git remote; Windows symlinks need Developer Mode |
Where this loses: setup overhead (one-time), the remote dependency, and Windows symlink ergonomics.
Where it wins: a single source for the whole agent surface (skills, subagents, MCP, OpenSpec scaffolding), refreshed with one git fetch.
The specialised subagents live as one canonical Markdown file each in subagents/ (this repo's root, not shipped to consumer projects). A generator emits the two tool-specific copies that are shipped.
After editing any canonical file:
python tools/gen_subagents.pyCI gate (fails if generated tree is stale relative to canonical):
python tools/gen_subagents.py --checkRules:
Kilo directory.
Adding a new subagent: drop a frontmatter-headed Markdown file in subagents/ and re-run the generator.
Removing one: delete the canonical file and re-run. Orphaned outputs are pruned automatically.
Updates are driven from docs/AGENTS-UPDATE.md, a doc shipped from upstream that refreshes itself on every run. It contains bash and PowerShell commands that:
docs/AGENT_TOOLING.md, docs/MCP_SETUP.md, and docs/AGENTS-UPDATE.md itself)..agents/skills/ and pull only those (no surprise additions)..claude/agents/ and .opencode/agents/ and pull only those.Files intentionally NOT touched by the update: the symlinked skill directories (.claude/skills/, .opencode/skills/, .codex/skills/), .claude/CLAUDE.md, AGENTS.md.example, the .kilocode/*.example templates, opencode.json.example, .mcp.json.example, and the consumer's customised AGENTS.md.
If your project was imported before this doc shipped and has no docs/AGENTS-UPDATE.md, pull it with a one-off git checkout agent-standards/master -- docs/AGENTS-UPDATE.md, then use it for every later refresh.
Full walkthrough lives in docs/AGENT_TOOLING.md.
Naming convention: UPPERCASE filenames ship to consumer projects via the git checkout above; lowercase filenames stay in this repo only.
Shipped to consumer projects:
| File | What's in it |
|---|---|
| docs/AGENT_TOOLING.md | Initial import flow, future updates, subagents, MCP overview, full OpenSpec integration and workflow. |
| docs/MCP_SETUP.md | Prerequisites, key acquisition, env-var export per OS, Claude Desktop and Codex CLI global configs, verification. |
| docs/AGENTS-UPDATE.md | Per-OS selective update commands. Refreshes the shipped docs, plus the skills and subagents already present, and itself. |
This repo only:
| File | What's in it |
|---|---|
| docs/repository-layout.md | Repo structure, agent compatibility matrix, instruction precedence per agent. |
| docs/manual-setup.md | Fallback steps when the standard git checkout flow needs manual help (symlink recreation per OS) plus per-agent start commands. |
| docs/bootstrap-prompt.md | First-run verification prompt to paste into the agent after import. Checks skills, subagents, MCP runtime, OpenSpec, instruction tree. |
Template:
| File | What's in it |
|---|---|
| AGENTS.md.example | Source instructions template. Copy to AGENTS.md after import; consumer projects fill the project-specific sections. |
Issues and pull requests welcome. For non-trivial changes please open an issue first to discuss the approach. The repo is opinionated by design and direction matters.
Maintained by Łukasz Sarna.
MIT, see LICENSE.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.