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/mmacpherson/tenforty/agents-mdgit clone --depth 1 https://github.com/mmacpherson/tenfortyWrote 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/mmacpherson/tenforty/agents-md)<a href="https://agentmods.dev/instructions/mmacpherson/tenforty/agents-md"><img src="https://agentmods.dev/badge/instructions/mmacpherson/tenforty/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 5.1 | $0.02685 | $0.02685 |
| Opus 5 | $0.01342 | $0.01342 |
| Sonnet 5 | $0.00537 | $0.00537 |
| Haiku 4.5 | $0.00268 | $0.00268 |
Grade A, and why
tenforty 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 4d 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 — 207 lines — stays where its author put it; the contents beside it link to each section on GitHub.
tenforty — Contributor & Agent Guide
Guidance for anyone working in this repo, human or AI coding agent. This is the
canonical guide; CLAUDE.md is a symlink to this file so tool-specific loaders
pick up the same content. It follows the AGENTS.md convention.
Working in this repo
mainis protected — changes land through a pull request, not direct pushes.- Branch for each change, and keep PRs focused.
- Never force-push
mainor any branch with an open PR. - Run the relevant quality gates before requesting review.
Vendored OpenTaxSolver is not ours to edit
We vendor OpenTaxSolver, we do not fork it. Never change what OTS computes —
not the release tarballs, not the generated ots_amalgamation.cpp, and not
through a patch function in ots/amalgamate.py. This holds even when the defect
is proven and the fix is one character.
When you find a defect in OTS's tax logic:
- Report it upstream. Stage the report in
docs/upstream-ots-reports.md. - Record it locally as a strict-xfail test plus a known-defect signature in
tests/taxcalc/taxcalc_policy.py. The xfail is the durable record and flips on its own once a release carries the correction.
Patching the vendored source instead would fork it invisibly: our tree would silently diverge from the upstream we claim to wrap, with nothing in the OTS release to show for it.
The narrow exception is portability and memory safety — making the source
compile and not read out of bounds, without changing any computed figure. The
existing patch functions are of this kind: C99→C++ shims, and
patch_az_widow_std_deduction for an out-of-bounds array read. Anything in this
category still gets reported upstream, and the patch function carries a docstring
saying why it qualifies.
Before deciding a finding is upstream's, check whether it is actually ours:
the mapping layer (models.py input maps, core.py activation and
orchestration) is our code and gets fixed here.
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.
- 4d ago First seen · 207 lines · 2,685 tokens per session scan A 4354a61e87bf
tenforty AGENTS.md is an instructions file published in the GitHub repository mmacpherson/tenforty (84 stars, last pushed 12d ago), licensed MIT. It adds 2,685 tokens to every session, about $0.0134 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 2026-09-01.
Other instructions, from other repositories
ZATCA AGENTS.md
AGENTS.md instructions for SallaApp/ZATCA: See CLAUDE.md for the full guide. Quick facts.
ayunis-legal-mcp CLAUDE.md
Instructions for ayunis-core/ayunis-legal-mcp, covering legal mcp - architecture and development guide, project overview, architecture, directory structure and key components.
korean-firefighter-law-mcp CLAUDE.md
Claude Code instructions for ssd7830-cmyk/korean-firefighter-law-mcp, covering korean-firefighter-law-mcp, a. 답변 규칙 — 기억이 아니라 이 저장소의 코드가 근거다, b. mcp 도구 사용 and c. 작업 규칙.
uae-regulations-registry AGENTS.md
AGENTS.md instructions for nshield-security/uae-regulations-registry, covering agents.md — instructions for ai agents, what this repo contains, preferred entry points, json schema (data/regulations.json) and stable urls.
spain-ai-kit CLAUDE.md
Instructions for aplaceforallmystuff/spain-ai-kit, a project described as: MCP servers connecting AI applications to Spanish government open data and legal infrastructure.
law-mcp CLAUDE.md
Claude Code instructions for finalchild/law-mcp, covering claude.md, build and development commands, architecture overview, core components and key integration points.