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/yakew7/fair-code/claude-mdgit clone --depth 1 https://github.com/yakew7/Fair-CodeWhat 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 | $0.00477 | $0.00477 |
| Opus 5 | $0.00238 | $0.00238 |
| Sonnet 5 | $0.00095 | $0.00095 |
| Haiku 4.5 | $0.00048 | $0.00048 |
Grade A, and why
Fair-Code CLAUDE.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 — 35 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md - Standing Instructions
Read this before making changes to this repo. It applies to AI agents and human contributors alike.
Paper freeze - lifted
This repo was previously under a paper freeze, on the assumption that a research paper citing its
benchmark results was in peer review. That paper was never actually submitted this cycle - the real
paper, with a fresh run of results, is now planned for next year. Freezing analysis code and
results against a paper that doesn't exist yet was premature, so the freeze is lifted: faircode/
core, results/, audit.yaml manifests, dataset CSVs, requirements-lock.txt, and the reproducibility
parameters (random_state, split ratio, iteration counts, the fairness constraint, the six metrics)
are all open to normal development again. New audits merge to main like anything else. Dependabot's
version-bump PRs against requirements-lock.txt no longer need to be reverted on sight - use normal
judgment.
paper/results-frozen/ stays, as a reference
paper/results-frozen/ is kept in the repo, but it is no longer under any special protection - it's
a snapshot of an earlier analysis pass, useful as a point of comparison while building toward next
year's real paper run, not the evidence backing a live citation. Nothing stops it from being read,
compared against, or eventually superseded; there's just no active enforcement requiring results/
to keep matching it, since results/ is expected to move again as development resumes. The
v1.0-paper git tag is left in place as a historical marker of what that snapshot was.
If an explainer or piece of documentation quotes a Fair Code benchmark result, say which source it's
quoting - results/ (current) or paper/results-frozen/ (the earlier reference snapshot) - so a
reader isn't left guessing which numbers are live.
Next freeze
When the real paper submission is actually ready (expected next year), expect a new freeze to be established at that point, covering whatever fresh results back that submission - this file will be updated again when that happens. Until then, treat this repo as under normal, fully open development.
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 · 35 lines · 477 tokens per session scan A 58bfa04ae39b
Fair-Code CLAUDE.md is an instructions file published in the GitHub repository yakew7/Fair-Code (48 stars, last pushed 2d ago), licensed MIT. It adds 477 tokens to every session, about $0.0024 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 2026-08-30.
Other instructions, from other repositories
awesome-trustworthy-deep-learning AGENTS.md
Instructions for MinghuiChen43/awesome-trustworthy-deep-learning, covering agents.md, project context, default task: add an arxiv paper, 1. normalize the arxiv id and 2. fetch metadata reproducibly.
HealthChain CLAUDE.md
Instructions for healthchainai/HealthChain, covering healthchain - claude code context, 0. project overview, 1. non-negotiable golden rules, 2. testing discipline and 3. build, test & utility commands.
medical-terminologies-mcp CLAUDE.md
Instructions for SidneyBissoli/medical-terminologies-mcp, covering claude.md, commands, runtime requirements, architecture and two entry points, one registration module (sdk v2).
HealthClawGuardrails GEMINI.md
Instructions for aks129/HealthClawGuardrails, covering healthclaw guardrails — gemini extension context, what the tools do and how to behave.
healthcareskills AGENTS.md
AGENTS.md instructions for aks-builds/healthcareskills, covering agents.md, repository overview, repository structure, agent skills specification and required frontmatter.
ai4science-studio CLAUDE.md
Instructions for AMDResearch/ai4science-studio, covering claude.md, repository purpose, agent entry points (read these first), directory layout and model slug rule.