Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/youwei792/agentsmd/agents-mdgit clone --depth 1 https://github.com/youwei792/agentsmdWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00521 | $0.00521 |
| Opus 5 | $0.00260 | $0.00260 |
| Sonnet 5 | $0.00104 | $0.00104 |
| Haiku 4.5 | $0.00052 | $0.00052 |
Grade A, and why
agentsmd AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 74 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI coding agents working in this repository.
agentsmd is a zero-dependency Go CLI that keeps AI agent instruction files (AGENTS.md, CLAUDE.md, GEMINI.md) healthy: it generates a grounded AGENTS.md from repository facts, validates every command and file reference, audits quality and token cost, and bridges CLAUDE.md/GEMINI.md to AGENTS.md.
Commands
All commands run from the repository root.
Build
go build -o agentsmd .
Testing
go test ./...
Lint & static analysis
go vet ./...
gofmt -l .
Try it end to end
go run . doctor .
go run . analyze .
Code style
- Formatting is enforced by gofmt: run
gofmt -won touched files. - Standard library only — no third-party dependencies. Detection that needs a parser (package.json is JSON) uses encoding/json; TOML/YAML uses the pragmatic line readers in internal/analyze/tomlish.go.
- Keep validation conservative: when the engine is not certain a reference
is broken, it must stay silent. False positives destroy trust in
check. - Errors go to stderr with exit code 1; findings are data (see validate.Finding).
Testing notes
Tests are table-driven and build fixture repositories in t.TempDir(); nothing reads the network or the host filesystem outside the fixture.
- internal/analyze_test.go — toolchain detection per ecosystem
- internal/mdutil_test.go — markdown command/path extraction
- internal/validate_test.go — the conservative-checking contract
- internal/cli_test.go — end-to-end: builds the binary and runs init → sync → doctor
Run a single package with go test ./internal/validate/.
Gotchas & conventions
- This tool parses its own AGENTS.md in CI (the dogfood job) — every command
documented here must actually work,
go run .form included. - Agentsmd supports flags after positional arguments (e.g.
agentsmd tokens . --json) via normalizeArgs in internal/cli/cli.go; keep new flags listed in boolFlags/valueFlags or they will not parse. - Windows path handling: always filepath.Join / filepath.ToSlash, never "/".
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 74 lines · 521 tokens per session scan A d467c00d1098
agentsmd AGENTS.md is an instructions file published in the GitHub repository youwei792/agentsmd (1 stars, last pushed yesterday), licensed MIT. It adds 521 tokens to every session, about $0.0026 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
agents-md-sync AGENTS.md
Instructions for trick77/agents-md-sync, covering agent instructions — agents-md-sync, project, stack, coding and testing.
agentlint AGENTS.md
AGENTS.md instructions for agentlint/agentlint, covering agents.md — agentlint, project, build, test, lint, typecheck, architecture and conventions.
agentlint CLAUDE.md
Claude Code instructions for agentlint/agentlint, covering claude.md — entry point for agents working on this repo, reading order (do not skip), operating principles (the short version), closing-session ritual and co-authorship.
agentsmd AGENTS.md
AGENTS.md instructions for daichunghy/agentsmd, covering agents.md, setup, build, test and conventions.
agents-md-doctor AGENTS.md
Instructions for ItsHege/agents-md-doctor, covering agents.md doctor public agent instructions, scope of this file, project goal, current baseline and working rules.
agentlint copilot-instructions.md
Copilot instructions for agentlint/agentlint, a project described as: The Lighthouse for AI coding agents. Audit any repo for how ready it is for Claude Code, Cursor, Codex, Copilot, and Gemini CLI. 0–100 score in 30 seconds.