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/zts0hg/codexspec/agents-mdgit clone --depth 1 https://github.com/Zts0hg/codexspecWhat 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 | $0.00823 | $0.00823 |
| Opus 5 | $0.00411 | $0.00411 |
| Sonnet 5 | $0.00165 | $0.00165 |
| Haiku 4.5 | $0.00082 | $0.00082 |
Grade A, and why
codexspec 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.
How it starts
The opening of the file, as written. The whole thing — 48 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
CodexSpec
This project uses CodexSpec for requirements-first spec-driven development.
Use these Codex skills when working on CodexSpec workflows:
$codexspec:constitutionto create or update project principles.$codexspec:specifyto capture confirmed requirements.$codexspec:generate-specto producespec.md.$codexspec:spec-to-designto producedesign.md.$codexspec:spec-to-planto produceplan.md.$codexspec:plan-to-tasksto producetasks.md.$codexspec:implement-tasksto implement approved tasks.$codexspec:distillto capture reusable cross-feature knowledge into.codexspec/profile/.$codexspec:evolveto contribute vetted profile knowledge back upstream via a reviewed PR.
Before making workflow decisions, read .codexspec/memory/constitution.md.
Maintainer Guide (developing CodexSpec itself)
This section is hand-authored (outside the managed block above) and applies whether you use Codex or Claude Code — the two context files are equal-status entry points.
The full maintainer development guide lives in CLAUDE.md: architecture and per-feature design notes, the self-bootstrap rule (edit templates/ and src/codexspec/, never the derived .claude/commands/ or .agents/skills/), and the two-constitutions distinction. It imports the project constitution (.codexspec/memory/constitution.md) and the repository layout / packaging boundary (docs/internal/repository-layout.md).
Read CLAUDE.md and docs/internal/repository-layout.md before changing templates, packaging, or init behavior.
CodexSpec Project Profile
Project constraints (highest priority — read these FIRST): before any non-trivial work you MUST read every record under .codexspec/profile/constraints/ — the project's hard prohibitions (严禁 / 仅允许). Honor them before anything else.
Actively recall relevant knowledge before non-trivial work. Do not wait to stumble on it: scan the record headings and their scope/when (and, for strategies, trigger) lines across the categories below, and read every record whose condition matches the task you are about to start (each directory holds one record per file):
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 · 48 lines · 823 tokens per session scan A 12f67ff395c4
codexspec AGENTS.md is an instructions file published in the GitHub repository Zts0hg/codexspec (5 stars, last pushed 3d ago), licensed MIT. It adds 823 tokens to every session, about $0.0041 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.
Other instructions, from other repositories
intlayer CLAUDE.md
Instructions for aymericzip/intlayer, covering repo layout, commands (run from root), content declaration flow, git and commit conventions.
flarestarter AGENTS.md
Instructions for FlareStarter/flarestarter, covering agents.md, what this is, structure, conventions and commands.
AnvilWiki AGENTS.md
Instructions for PNGTRID/AnvilWiki, covering agents.md, repository purpose, read these first, intended tech stack (verified, as of 2026-08-11) and architecture: code/config/content separation (critical).
txtify CLAUDE.md
Instructions for lkmeta/txtify, covering txtify — agent guide, architecture (read this before touching job flow), commands, invariants & known traps and conventions.
claude-code-zh-cn AGENTS.md
Instructions for taekchef/claude-code-zh-cn, covering claude-code-zh-cn, 项目结构, 数据流, 技术要点 and 维护流程规则.
tolgee-platform AGENTS.md
AGENTS.md instructions for tolgee/tolgee-platform, covering agents.md, repository structure, backend development, database migrations and backend testing.