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/whenpoem/aiscientist/agents-mdgit clone --depth 1 https://github.com/whenpoem/aiscientistWrote 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/whenpoem/aiscientist/agents-md)<a href="https://agentmods.dev/instructions/whenpoem/aiscientist/agents-md"><img src="https://agentmods.dev/badge/instructions/whenpoem/aiscientist/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 | $0.01865 | $0.01865 |
| Opus 5 | $0.00932 | $0.00932 |
| Sonnet 5 | $0.00373 | $0.00373 |
| Haiku 4.5 | $0.00186 | $0.00186 |
Grade A, and why
aiscientist 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 — 114 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
中文版本: AGENTS.zh-CN.md Operating instructions for any later LLM or coding agent working in this repository. For project orientation and architecture, read
README.mdanddocs/overview.mdfirst — this file assumes you already know what the project does.
Editing rules
- Prefer small, explicit changes over broad rewrites. This repository is small enough that rewrites are tempting and almost always wrong.
- Keep Windows compatibility. The target machine is Windows 11. Prefer ASCII unless the file already requires otherwise.
- If you change behavior, update or add tests in the same area. Behavioral changes without test updates are regressions waiting to happen.
- Do not claim a fix works until you rerun the relevant command. "Should work" is not the same as "works".
- Do not say "production-ready" based only on unit tests. Unit-test green is necessary, not sufficient.
- Do not modify files in
.research-agent/. That directory is runtime state, not source. Use the MCP tools to mutate it. - Do not bypass
query_heldoutto read sequestered data. The leakage hook will block you anyway, and bypassing it leaves an audit gap.
Verification rules
Before claiming success, run the checks that match the change.
Minimum baseline:
uv run ruff check
uv run pytest tests/memory_mcp tests/verify_mcp tests/prove_mcp tests/hooks tests/cockpit tests/scripts tests/e2e
If cockpit code changed, also smoke-test the TUI entrypoint:
uv run python -m cockpit.tui --once
uv run python -m cockpit.tui --once --lang zh
If integration points changed, also smoke-test the backend:
uv run python -c "import memory_mcp.server; import verify_mcp.server; import prove_mcp.server; import cockpit.mcp_server; print('OK')"
What to read before a non-trivial change
In this order:
README.md— orientationdocs/overview.md— mental modeldocs/architecture.md— cross-module contracts (treat as binding).claude/settings.json— MCP and hook wiring- The relevant package under
src/ - The matching tests under
tests/ docs/archive/only if the change is about historical scope
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 · 114 lines · 1,865 tokens per session scan A db1a1bab40d5
aiscientist AGENTS.md is an instructions file published in the GitHub repository whenpoem/aiscientist (8 stars, last pushed 1mo ago), licensed MIT. It adds 1,865 tokens to every session, about $0.0093 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
codex-autoresearch AGENTS.md
Instructions for TheGreenCedar/codex-autoresearch, covering agents.md, purpose and scope, canonical sources, stable product boundaries and source and package boundaries.
autoharness AGENTS.md
Instructions for tigerless-labs/autoharness: This repo uses Open Knowledge — collaborative markdown via MCP. .open-knowledge/config.yml (with optional /.open-knowledge/config.yml; CLI/env may override) is the path contract: content.dir is the root for relative paths; content.include lists globs that add markdown…
MindFlock AGENTS.md
Instructions for MindFlock/MindFlock, covering mindflock — agent onboarding, setup and tests, map and conventions.
opencode-dejavu AGENTS.md
Instructions for WhiteBite/opencode-dejavu, covering project knowledge base, structure, where to look, code map and conventions.
Polaris AGENTS.md
Instructions for sponge-b0b/Polaris, covering agents.md, purpose and authority, domain vocabulary, coding conduct and secrets.
nbp-skillforge AGENTS.md
Instructions for nbpadilha/nbp-skillforge, covering agents.md, what this is, the one rule that matters, before you finish a change and conventions.