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/agentlint/agentlint/agents-mdgit clone --depth 1 https://github.com/agentlint/agentlintWhat 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 | $0.02439 | $0.02439 |
| Opus 5 | $0.01220 | $0.01220 |
| Sonnet 5 | $0.00488 | $0.00488 |
| Haiku 4.5 | $0.00244 | $0.00244 |
Grade A, and why
agentlint 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 — 164 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — agentlint
Briefing for AI coding agents working in this repo. Concise and actionable. Source of truth that all other tool-specific files (CLAUDE.md, .cursor/rules, copilot-instructions.md) point at.
Project
agentlint is a TypeScript monorepo using pnpm workspaces. It ships an open-source CLI (packages/cli) that scans a repository against ~30 checks and produces a 0–100 agent-readiness score, plus a core package with pure types and score calculation.
Build, test, lint, typecheck
All commands run from the repo root. Package manager: pnpm (>=9).
pnpm install # install deps for all workspaces
pnpm run build # build everything
pnpm run test # run tests (vitest) across all workspaces
pnpm run lint # biome check
pnpm run typecheck # tsc --noEmit across the workspace
pnpm run agentlint . # run agentlint on this repo (must score 100)
pnpm run ci # full pipeline: build + typecheck + lint + test + agentlint
Always run pnpm run ci before considering work done. CI runs the same pipeline; if the score drops below 100, the build fails.
Architecture
packages/
core/ pure types, score-calc — no IO
src/types.ts Result, Rule, Report, ScanContext
src/score.ts buildReport, registerRuleCategory
cli/ CLI entrypoint, FS walker, reporters
src/index.ts bin entry — argv parsing + subcommand dispatch
src/api.ts programmatic `runScan` export (package `main`)
src/init/index.ts `agentlint init` subcommand (v2)
src/prompt-cmd/ `agentlint prompt` subcommand (consolidated fix prompt)
src/prompts/ predefined fix-prompt templates (registry) + composer
src/scan-context.ts cached FS reads + project meta detection
src/push/ push pipeline
client.ts POST /api/runs with project token
config.ts .agentlint.json loader (ADR-0020)
oidc.ts GitHub Actions OIDC token fetcher (ADR-0019)
project-lookup.ts GET /api/cli/projects (used by init)
token.ts AGENTLINT_TOKEN env resolution
pr-detect.ts PR context detection from CI env
repo-detect.ts git remote → owner/name
src/rules/ one file per category, six rules each
src/report/ terminal, html, json, markdown reporters
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 · 164 lines · 2,439 tokens per session scan A bd14516cce90
agentlint AGENTS.md is an instructions file published in the GitHub repository agentlint/agentlint (1 stars, last pushed 8d ago), licensed MIT. It adds 2,439 tokens to every session, about $0.0122 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
fastagent AGENTS.md
Instructions for fastagent-sh/fastagent, covering fastagent — agent guide, what this is, source of truth, repo map and devx principle stack.
agents-reverse-engineer AGENTS.md
Instructions for GeoloeG-IsT/agents-reverse-engineer, covering agents-reverse-engineer, stack, subdirectories, architecture and behavioral contracts.
agent-operating-protocol AGENTS.md
Instructions for ryu-tada/agent-operating-protocol, covering agents.md - agent operating protocol, 0. instruction priority and trust boundaries, 1. core operating stance, 2. why high-agency agents feel strong and 3. default work loop.
agentsmd AGENTS.md
AGENTS.md instructions for youwei792/agentsmd, covering agents.md, commands, build, testing and lint & static analysis.
agents-md-facts AGENTS.md
Instructions for Wolfe-Jam/agents-md-facts, covering agents.md — agents-md-facts, setup & build, run the tests, where things live and conventions.
agents-reverse-engineer CLAUDE.md
Instructions for GeoloeG-IsT/agents-reverse-engineer, a project described as: Reverse engineer your codebase to let your agents work efficiently.