agentmods.dev

agent47 AGENTS.md

Repository instructions for the agent47 project, including required reading, working rules, and the agreement between coding agents. An SDK is a software package that developers use to build with another system.

In plain words
What is it for?
Use them when working on the repository's SDK, checking its memory files, following its free and MIT-licensed requirements, or recording merged changes for the cofounder agent.
Why use it?
They give agents the project's current decisions and constraints before they change code, reducing contradictions and missed handoffs.

Instructions file for CodexOpenCode

4 3d ago A scan 4,862 tokens original MIT
View the source file bmdhodl/agent47
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/agents-md
Clone the repo
git clone --depth 1 https://github.com/bmdhodl/agent47

Made for: Codex, OpenCode.

Per session 4,862 This file is loaded in full into every session.
When invoked 4,862 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 $0.04862 $0.04862
Opus 5 $0.02431 $0.02431
Sonnet 5 $0.00972 $0.00972
Haiku 4.5 $0.00486 $0.00486

Measured 2d ago against content hash db88764bb8a4, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent47 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 · 411 lines

How it starts

The opening of the file, as written. The whole thing — 411 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

READ THIS FIRST: SDK Memory

Before touching any code, read these tracked SDK-only memory files. They are the ground truth for this public repository.

  • memory/state.md — SDK version, packaging state, current focus
  • memory/blockers.md — what is blocked, deferred, or not worth more cycles
  • memory/decisions.md — locked SDK product decisions. Do NOT contradict.
  • memory/distribution.md — positioning, channels, and audience for SDK growth

Updated regularly. If memory/ conflicts with older info in this file, memory/ wins.

Cofounder Inbox

Use inbox/ to communicate concise SDK status to the cofounder agent.

  • inbox/log.md is the durable handoff file.
  • Write one entry after each merged PR.
  • Format each entry as: date, agent, what shipped, decisions made, blockers.
  • Keep it SDK-only and short.
  • Do not dump business-sensitive plans there.

Key constraints:

  • SDK stays free, MIT, zero-dependency. This is non-negotiable.
  • Do not build paid features in this repo.
  • Distribution > features. Get listed on registries before building new guards.
  • Focus is runtime enforcement + coding-agent safety.
  • Do not store business-sensitive plans or outreach data in this repo.

This file provides guidance to Codex (Codex.ai/code) when working with code in this repository.

What This Is

AgentGuard — a zero-dependency runtime guardrails SDK for coding agents and AI agents. The SDK is open source (MIT, zero dependencies).

  • Repo: github.com/bmdhodl/agent47
  • Dashboard repo: github.com/bmdhodl/agent47-dashboard (private)
  • Package: agentguard47 on PyPI (release status tracked in memory/state.md; release candidate: v1.2.14)
  • Landing page: site/index.html (Vercel)

Agent Contract (MANDATORY)

Every Codex session in this repo MUST follow these rules. No exceptions.

  1. Read ops/ docs first. Before starting any task, read ops/00-NORTHSTAR.md, ops/03-ROADMAP_NOW_NEXT_LATER.md, and ops/04-DEFINITION_OF_DONE.md.
  2. Stop on conflict. If your task conflicts with NORTHSTAR or ROADMAP, stop and propose a correction — do not proceed.
  3. Restate before coding. Before writing any code, explicitly restate:
    • (a) Goal — what are we achieving?
    • (b) Scope — what files/modules are touched?
    • (c) Non-goals — what are we NOT doing?
    • (d) Done criteria — how do we know it's done? (reference ops/04-DEFINITION_OF_DONE.md)
  4. Structured output. Every task must include: plan → diff summary → tests → docs updates needed.
  5. Check staleness. Run git log -1 --format='%cr' -- ops/03-ROADMAP_NOW_NEXT_LATER.md and git log -1 --format='%cr' -- ops/02-ARCHITECTURE.md. If ROADMAP is >5 days old or ARCHITECTURE is >14 days old, warn the user before starting any task.
  6. PR proof is required. Every PR must include concrete proof that the change works: command output, targeted runtime evidence, screenshots when applicable, or saved artifacts under a local proof folder.
  7. Always do the post-PR review loop. After opening a PR: wait for CI, verify the relevant preview/deployment health, wait a few minutes for automated review, inspect the full PR timeline plus review comments/threads, address feedback, rerun checks, and only then call the PR ready.
  8. Use matching built-in skills by default. When the task matches them, use playwright for browser automation and screenshot proof, playwright-interactive when persistent browser state helps, gh-address-comments for PR review/comment sweeps, and vercel-deploy for deployment work. Do not skip these when the task clearly fits.
  9. Write only high-signal inbox updates. If the cofounder agent should know about a material SDK blocker or milestone, append one concise entry to inbox/log.md after the merged PR instead of writing a long narrative.

Read the full file on GitHub · 411 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 · 411 lines · 4,862 tokens per session scan A db88764bb8a4

Subscribe to this mod's changes

agent47 AGENTS.md is an instructions file published in the GitHub repository bmdhodl/agent47 (4 stars, last pushed 3d ago), licensed MIT. It adds 4,862 tokens to every session, about $0.0243 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.

AI SafetyGuardrailsLangChainMemory and contextPlanningai-agentsanthropicdeveloper-tools
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