agent47 CLAUDE.md

Repository instructions for AgentGuard, whose public code includes a zero-dependency Python safety SDK and a read-only MCP server.

In plain words
What is it for?
Guiding coding agents through AgentGuard changes while preserving its free MIT SDK, runtime safety focus, repository boundaries, and required definition of done.
Why use it?
They define which files to read, what belongs in the repository, and which product or business areas are outside its scope.

Instructions file

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/bmdhodl/agent47/claude-md
Clone the repo
git clone --depth 1 https://github.com/bmdhodl/agent47
Per session 1,283 This file is loaded in full into every session.
When invoked 1,283 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.

ModelPer sessionOnce invoked
Fable 5 $0.01283 $0.01283
Opus 5 $0.00642 $0.00642
Sonnet 5 $0.00257 $0.00257
Haiku 4.5 $0.00128 $0.00128

Measured 2d ago against content hash 34763bd2280d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent47 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 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.

CLAUDE.md · 153 lines

How it starts

The opening of the file, as written. The whole thing — 153 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

Claude Code already injects broad tool, style, memory, and session guidance. This file is only for AgentGuard-specific repo rules that Claude would not know otherwise.

Read First

Before touching code or docs, read these files in this order:

  1. memory/state.md
  2. memory/blockers.md
  3. memory/decisions.md
  4. memory/distribution.md
  5. ops/00-NORTHSTAR.md
  6. ops/03-ROADMAP_NOW_NEXT_LATER.md
  7. ops/04-DEFINITION_OF_DONE.md
  8. ARCHITECTURE.md

If memory/ conflicts with older repo docs, memory/ wins.

Repo Boundary

AgentGuard is the public SDK wedge.

  • sdk/ = zero-dependency Python runtime guardrails SDK
  • mcp-server/ = narrow read-only MCP surface
  • hosted dashboard = private repo, out of scope here

Non-negotiable:

  • SDK stays free, MIT, and zero-dependency
  • distribution beats random feature expansion
  • focus stays on runtime enforcement + coding-agent safety
  • do not add paid/dashboard features in this repo
  • do not put business-sensitive plans or outreach data in this repo

Claude Repo Contract

Before coding, explicitly restate:

  • goal
  • scope
  • non-goals
  • done criteria tied to ops/04-DEFINITION_OF_DONE.md

If ops/03-ROADMAP_NOW_NEXT_LATER.md is older than 5 days or ARCHITECTURE.md is older than 14 days, warn before starting work.

Always:

  • read before editing
  • keep changes minimal and product-facing
  • stop if the request conflicts with memory/ or ops/
  • leave proof for every PR
  • do the post-PR review loop: CI, automated review, comment sweep, fixes, rerun

What Claude Should Optimize For Here

  • onboarding and activation
  • docs, quickstarts, and starter quality
  • SDK hardening and tests
  • eval/report/incident proof surfaces
  • release-readiness and metadata consistency
  • distribution-supporting assets tied to the actual SDK and MCP product

Avoid:

  • dashboard-only work
  • speculative framework sprawl
  • broad observability positioning
  • novelty features with weak adoption value
  • large refactors without user-facing payoff

Read the full file on GitHub · 153 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.

  1. 2d ago First seen · 153 lines · 1,283 tokens per session scan A 34763bd2280d

Subscribe to this mod's changes

agent47 CLAUDE.md is an instructions file published in the GitHub repository bmdhodl/agent47 (4 stars, last pushed 4d ago), licensed MIT. It adds 1,283 tokens to every session, about $0.0064 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.

Related

Other instructions, from other repositories

Doberman-Core AGENTS.md

Instructions for DobermanCore/Doberman-Core, covering claude.md — doberman operating manual, 0. on startup (every session), 1. what this repository is, 2. architecture & extension points and 3. prime directives (non-negotiable).

DobermanCore/Doberman-Core · 3,547 tokens

FeatherlaneAI AGENTS.md

Instructions for ducnguyen67201/FeatherlaneAI, covering repository agent instructions, skills, architecture: rust backend is the source of truth, page integration expectations and implementation checklist.

ducnguyen67201/FeatherlaneAI · 4,372 tokens

selectools CLAUDE.md

Instructions for johnnichev/selectools, covering selectools — agent instructions, commands, stability markers, steptype reference (27 types) and common pitfalls (from past bugs).

johnnichev/selectools · 1,672 tokens

selectools AGENTS.md

Instructions for johnnichev/selectools, covering selectools -- agent instructions, commands, test (must pass before any commit), type check and security (must pass before release tags).

johnnichev/selectools · 1,252 tokens

taintgate AGENTS.md

Instructions for Nikita3005/taintgate, covering taintgate engineering guide, core principle, engineering priorities, python and security.

Nikita3005/taintgate · 259 tokens

AUTHENSOR AGENTS.md

Instructions for AUTHENSOR/AUTHENSOR, covering agents.md — authensor integration guide for ai agents, what authensor does for you, quick integration, if you have access to http and if you have access to npm / npx.

AUTHENSOR/AUTHENSOR · 1,298 tokens