What is burrow AGENTS.md?

A project-specific instruction file for Burrow, a command-line runtime that runs coding agents inside operating-system sandboxes with limited file access.

Use it when developing or testing Burrow, its command-line tools, sandbox behavior, sessions, or event storage.

It explains the security purpose, system design, commands, and quality checks agents must follow in the repository.

burrow AGENTS.md — Instructions file by jayminwest

Install

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.

agentmods

npx agentmods add instructions/jayminwest/burrow/agents-md

Clone the repo

git clone --depth 1 https://github.com/jayminwest/burrow

Made for: Codex, OpenCode.

Per session 1,792 This file is loaded in full into every session.

When invoked 1,792 The same file — it is already loaded in full.

Security scan A 0 findings. Scan, not verified.

Origin original No closer match found in the catalogue.

Token cost

What 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	$0.01792	$0.01792
Opus 5	$0.00896	$0.00896
Sonnet 5	$0.00358	$0.00358
Haiku 4.5	$0.00179	$0.00179

Measured 2d ago against content hash 041e9ba1086e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

burrow 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 2d ago.

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.

AGENTS.md · 187 lines

How it starts

The opening of the file, as written. The whole thing — 187 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

This file is the canonical entry point for AI coding agents working in burrow, following the agents.md convention.

Mission

burrow — OS-isolated sandbox runtime for coding agents.

Burrow wraps untrusted agent processes in OS-level sandboxes — bwrap on Linux, sandbox-exec on macOS — with no Docker and no daemon. Its single most important guarantee is that a sandboxed agent cannot read or write outside the filesystem scopes the operator granted it. Sessions and events persist in SQLite via bun:sqlite.

Burrow is part of the os-eco ecosystem: it is the substrate sandbox primitive that warren (the cloud control plane) embeds per run, and it works standalone via the burrow / bw CLI. SPEC.md is the frozen V1 design record; ROADMAP.md is the punch list of forward-looking R-NN items.

Commands

All commands run from the repo root. Bun must be on PATH.

bun install                   # install dependencies
bun test                      # run all tests
bun test src/foo.test.ts      # run a single test file
bun run lint                  # biome check --error-on-warnings .
bun run lint:fix              # biome check --write .
bun run typecheck             # tsc --noEmit
bun run check:all             # full quality-gate suite (see below)
bun run verify                # alias for check:all (agent-facing entry point)
bun run check:coverage        # tests + coverage-floor ratchet

The user-facing CLI entry points:

bw --help                     # CLI top-level help
bw up                         # start a sandboxed session

Quality gates

bun run check:all (quiet runner, scripts/check-all.ts) runs every gate CI also enforces, in canonical order:

lint — biome check --error-on-warnings . (warnings fail)
typecheck — tsc --noEmit
check:agents — scripts/validate-agents-md.ts (this file's references)
check:dups — bunx jscpd (duplicate-code detector, .jscpd.json)
check:deps — knip --dependencies (unused / undeclared deps, knip.json)
check:size — scripts/check-file-sizes.ts (line-count ratchet)
check:debt — scripts/check-debt-markers.ts (tracker-pinned TODOs)
check:coverage — scripts/check-coverage.ts (coverage-floor ratchet)
check:ci-parity — scripts/check-ci-parity.ts (CI ⇄ check:all parity)

Read the full file on GitHub · 187 lines

Changes

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.

2d ago First seen · 187 lines · 1,792 tokens per session scan A 041e9ba1086e

Subscribe to this mod's changes

burrow AGENTS.md is an instructions file published in the GitHub repository jayminwest/burrow (11 stars, last pushed 1mo ago), licensed MIT. It adds 1,792 tokens to every session, about $0.0090 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 2026-08-30.

ai-agents developer-tools sandbox

burrow AGENTS.md

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Nothing flagged

How it starts

AGENTS.md

Mission

Commands

Quality gates

What this file has done since we first saw it

Other instructions, from other repositories

OpenSandbox AGENTS.md

microsandbox AGENTS.md

agentos CLAUDE.md

moltis CLAUDE.md

OpenSandbox CLAUDE.md

agent-sandbox AGENTS.md