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/pwya/md-paper/agents-mdgit clone --depth 1 https://github.com/pwya/md-paperWrote 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/pwya/md-paper/agents-md)<a href="https://agentmods.dev/instructions/pwya/md-paper/agents-md"><img src="https://agentmods.dev/badge/instructions/pwya/md-paper/agents-md.svg" alt="Measured on agentmods" 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 | $0.00891 | $0.00891 |
| Opus 5 | $0.00445 | $0.00445 |
| Sonnet 5 | $0.00178 | $0.00178 |
| Haiku 4.5 | $0.00089 | $0.00089 |
Grade A, and why
md-paper 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 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 — 28 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md — md-paper rules for AI agents · 跨工具代理守则
中文速览:这是 md-paper 给【任何】AI 编程代理(Codex / OpenCode / Hermes Agent / Claude Code / …)的工作守则。Claude Code 有成熟保护钩子，Codex / OpenCode 可安装兼容适配层；无论钩子是否存在，下面铁律始终是第一层约束。
What this repo is
md-paper revises Word manuscripts through a Markdown source of truth (manuscript.md), in five stages — each a standard Agent Skill (<skill>/SKILL.md, agentskills.io-compatible):
md-unpack (docx → manuscript.md) → md-triage (any revision intent → human-confirmed checklist) → md-swarm (batch revision) / md-iterate (single-spot revision) → md-build (manuscript.md → Word with live Zotero fields).
- Opened this repo to install? Follow INSTALL.md step by step — it covers Claude Code, Codex, OpenCode, Hermes and generic agents.
- Working on a paper project that uses md-paper? The rules below are binding in every session.
Iron rules — binding even when a harness hook is unavailable or incomplete
- Never write
manuscript.mddirectly. Not with a file-edit tool, not with a script (open(...,'w'),Set-Content,Out-File, …). The ONLY writer ismd-swarm/apply_md_changeset.py(--dry-run→ apply → verify). In Claude Code a hook physically denies stray writes; in your harness nothing will stop you — so don't. (Measured incident: parallel direct writes left 1 surviving section and silently lost 38 citations.) - Citations are load-bearing. Never delete or rewrite
[@citekey]marks while editing prose — the apply gate refuses patches that drop citations; don't route around it. Never invent new keys that imitate existing ones: a manuscript full of[@authorYear]-looking keys may be an unreconciled provisional namespace, and imitation = fabrication. - The human gate is human-only.
swarm/md_triage.mdcarries the token**人工确认：** 待确认. Only the human flips it to已确认. Never flip or paraphrase it yourself. - Each stage's
SKILL.mdis the authoritative runbook. Read it in full before running that stage; call the pipeline scripts exactly as written there — don't improvise replacements. - Windows + PowerShell is the assumed environment (ingest needs Word COM). Encoding:
.py/.md= UTF-8 no BOM, LF;.ps1containing Chinese = UTF-8 with BOM.
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 First seen · 28 lines · 891 tokens per session scan A 8d777f6bce0c
md-paper AGENTS.md is an instructions file published in the GitHub repository pwya/md-paper (10 stars, last pushed 16d ago), licensed Apache-2.0. It adds 891 tokens to every session, about $0.0045 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
mq commit.instructions.md
Instructions for harehare/mq: Use the following format for commit messages.
mcp-pandoc AGENTS.md
Instructions for vivekVells/mcp-pandoc, covering agents.md, what this is, where things are, deeper context and commands.
research-os AGENTS.md
Instructions for lxinfei5/research-os, covering researchos — constitution, §0 what this is, §1 directory discipline, §2 half-life (innovation 1) — memory design and §3 behavior pillars (innovations 2–5).
markview CLAUDE.md
Instructions for paulhkang94/markview, covering markview — ai development guide, quick reference, architecture, mcp tools (for ai integration) and safety rules.
seekstone CLAUDE.md
Instructions for shaqmughal/seekstone, covering claude.md, what this repo is, commands, the harness itself (run after npm install) and architecture.
mnemonic memory-files.instructions.md
Instructions for modeled-information-format/mnemonic, covering memory file guidelines (mif level 3), required frontmatter fields, optional temporal fields, optional provenance fields and memory types.