Borrowing it
Nothing to install: this file belongs to XuebinMa/agent-guard. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/XuebinMa/agent-guard/main/CLAUDE.mdgit clone --depth 1 https://github.com/XuebinMa/agent-guardWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/xuebinma/agent-guard/claude-md)<a href="https://agentmods.dev/instructions/xuebinma/agent-guard/claude-md"><img src="https://agentmods.dev/badge/instructions/xuebinma/agent-guard/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/xuebinma/agent-guard/claude-md"><img src="https://agentmods.dev/badge/instructions/xuebinma/agent-guard/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 5.1 | $0.03719 | $0.03719 |
| Opus 5.5 | $0.01488 | $0.01488 |
| Sonnet 5.5 | $0.00744 | $0.00744 |
| Haiku 4.5 | $0.00372 | $0.00372 |
Grade A, and why
agent-guard CLAUDE.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 3d 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.
How it starts
The opening of the file, as written. The whole thing — 194 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Project Overview
agent-guard is an execution control layer for agent side effects. It sits between agent tool intent and real execution, evaluates calls against policy, applies validator and sandbox controls, and records auditable outcomes.
Today, the clearest proof point is structured recognition and policy decisions
for recognized outbound Git push forms. The existing shell/file/HTTP surfaces
remain supported, but new product work should not expand them horizontally. The
target broker boundary is: an agent may write and test freely; agent-guard
decides and executes which exact Git change may leave the machine.
Plumbing-level git send-pack is part of that same outbound boundary, not a
separate feature surface. Current source version: 0.2.10.
Build & Test Commands
# Canonical local verification entrypoint
./scripts/verify.sh full
# Rust-only verification path
./scripts/verify.sh rust
./scripts/verify.sh lint
# Build / test the Rust workspace without the PyO3 extension-module feature trap
cargo build --workspace --exclude agent-guard-python --all-features
cargo test --workspace --exclude agent-guard-python --all-features
# Test (single crate)
cargo test --package agent-guard-core
# Test (single test)
cargo test --package agent-guard-core --lib tests::test_policy_load
# Lint
cargo clippy --workspace --exclude agent-guard-python --all-features -- -D warnings
# Format check / fix
cargo fmt --all -- --check
cargo fmt --all
# Run an example
cargo run -p agent-guard-sdk --example quickstart
# Python binding (requires maturin; extension-module is opt-in)
cd crates/agent-guard-python && maturin develop --features extension-module
# Node binding
cd crates/agent-guard-node && npm run build
Workspace Architecture
Ten crates under crates/, layered bottom-up:
agent-guard-core ← foundational types, YAML policy engine, audit, attestation
↑
agent-guard-validators ← bash command + path validation (destructive detection, traversal checks)
agent-guard-sandbox ← multi-platform sandbox trait (seccomp, Seatbelt, Windows Job/AppContainer, noop)
↑
agent-guard-sdk ← main integration point: Guard struct, anomaly detection, metrics, provenance, SIEM
↑
agent-guard-python ← PyO3 bindings (maturin, abi3-py310)
agent-guard-node ← napi-rs bindings
guard-verify ← CLI: execution-receipt verification + host-boundary doctor
agent-guard-broker ← isolated, credential-bearing Git push transaction boundary
agent-guard-cli ← CLI: interactive approval workflow (bin: agent-guard)
guard-hook ← Claude Code PreToolUse hook adapter (bin: guard-hook)
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.
- 3d ago Changed · +1 lines · +26 tokens per session a19d6b965a8d
- 4d ago Changed · +2 lines · +36 tokens per session d73d6d5f50e5
- 5d ago Changed 3e96efaa4ed5
- 20d ago Changed · +16 lines · +198 tokens per session f362ffcfb52e
- 21d ago Changed · +11 tokens per session 0301ddb359f1
- 24d ago Changed · +5 tokens per session d397f4bfb63f
- 1mo ago Changed d63295fd6b8f
- 1mo ago Changed · +2 lines · +29 tokens per session b43f16a45289
- 1mo ago Changed · +23 lines · +917 tokens per session f7e85a1d7928
- 1mo ago First seen · 150 lines · 2,497 tokens per session scan A 1a34465324cb
agent-guard CLAUDE.md is an instructions file published in the GitHub repository XuebinMa/agent-guard (5 stars, last pushed 3d ago), licensed MIT. It adds 3,719 tokens to every session, about $0.0149 per session on Opus 5.5. 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
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode copilot-instructions.md
Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).
vscode model-prompts.instructions.md
Model-specific prompt authoring and registry guidelines.