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/kevinbigham/codex-safe-starter/agents-mdgit clone --depth 1 https://github.com/KevinBigham/codex-safe-starterWhat 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 | $0.01484 | $0.01484 |
| Opus 5 | $0.00742 | $0.00742 |
| Sonnet 5 | $0.00297 | $0.00297 |
| Haiku 4.5 | $0.00148 | $0.00148 |
Grade C, and why
codex-safe-starter AGENTS.md scanned grade C with 2 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.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
- Do not run unknown installers or `curl | sh` style commands. Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- Do not run unknown installers or `curl | sh` style commands. How it starts
The opening of the file, as written. The whole thing — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Project Purpose
This repository is Codex Safe Starter: a portable, Git-backed operating layer for safer Codex work. V3 packages canonical repo skills, project-scoped read-only agents, project-local config, setup checks, smoke validation, staged evals, portable CODEX_HOME templates, and reporting without mutating global Codex state.
Repo Map
MISSION.md: source mission; read before changing the stack.README.md: clone-anywhere quickstart.scripts/: bootstrap, doctor, smoke, secret scan, clean, and portable Codex launchers..agents/skills/: canonical repo-scoped Codex skills. Do not leave duplicate liveSKILL.mdfiles underskills/..codex/agents/*.toml: project-scoped read-only explorer/reviewer agents..codex-home-template/: optional portableCODEX_HOMEstarter copied into ignored.codex-home/..github/workflows/: smoke CI and optional Promptfoo eval CI.codex-upgrade/: baseline, findings, security review, evals, benchmarks, and proposed config/global artifacts.docs/codex/: durable planning, playbook, review, and testing guidance..codex/config.toml: project-local Codex defaults only. Do not add auth, provider, telemetry, notification, global profile, active MCP, active hook, active plugin, or secret settings here.
Do not casually touch files outside this repository. All Git operations must run from the isolated repo root.
Commands
- Install: no project install step.
- Dev server: none.
- Build: none.
- Bootstrap check:
./scripts/bootstrap.sh --check - Doctor:
node scripts/doctor.mjs - CI doctor:
node scripts/doctor.mjs --ci - Selftest:
node scripts/selftest.mjs - Smoke test:
node scripts/smoke.mjs - Secret scan:
node scripts/secret-scan.mjs - Legacy final smoke:
node codex-upgrade/benchmarks/run-smoke.mjs --final - Inspect instructions:
sed -n '1,220p' AGENTS.md - Search workspace:
rg -n "<term>" . - List workspace files:
find . -maxdepth 4 -type f | sort
If a command is missing because a future task removed or renamed the benchmark harness, update docs/codex/testing.md in the same change.
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 · 115 lines · 1,484 tokens per session scan C dd66c57ee745
codex-safe-starter AGENTS.md is an instructions file published in the GitHub repository KevinBigham/codex-safe-starter (1 stars, last pushed 1mo ago), licensed MIT. It adds 1,484 tokens to every session, about $0.0074 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
cookiecutter-mlops-package AGENTS.md
Instructions for fmind/cookiecutter-mlops-package, covering agents.md, project overview, setup & core commands, definition of done and conventions & idioms.
Codex-Skills AGENTS.md
Instructions for TheGoat395/Codex-Skills, covering repository operating notes, required checks, evidence rules and change hygiene.
karngyan.com CLAUDE.md
Instructions for karngyan/karngyan.com, covering claude.md, commands, architecture, conventions and recipes.
agent-skills AGENTS.md
Instructions for sayeed1999/agent-skills, covering agents.md, output style, output style, 1. think before code and 2. simplicity.
give-me-job AGENTS.md
Instructions for kyoungbinkim/give-me-job, covering give-me-job repository instructions, default workflow, skill routing, working on this repository and development loop.
nodejs-starter CLAUDE.md
Instructions for threeal/nodejs-starter, covering claude.md, about this repository, architecture, source files and build outputs.