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/taucetiproject/tauceti/agents-mdgit clone --depth 1 https://github.com/TauCetiProject/TauCetiWrote 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/taucetiproject/tauceti/agents-md)<a href="https://agentmods.dev/instructions/taucetiproject/tauceti/agents-md"><img src="https://agentmods.dev/badge/instructions/taucetiproject/tauceti/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 | $0.01483 | $0.01483 |
| Opus 5 | $0.00741 | $0.00741 |
| Sonnet 5 | $0.00297 | $0.00297 |
| Haiku 4.5 | $0.00148 | $0.00148 |
Grade A, and why
TauCeti 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 5d 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 — 99 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Working in Tau Ceti
This is the AI-owned code repo of Tau Ceti. Read README.md for what the project is and how
the three repos fit together; this file only adds the contract for agents working here.
Before you write code
Read the roadmap first. The roadmaps live in the separate TauCetiRoadmap repo. The roadmap gates new mathematics: only add a new mathematical declaration (definition, theorem, instance, notation) or file when it advances a specific roadmap target, or supplies a prerequisite that a specific target needs. If something you want to build is not on the roadmap — whether a human asked for it or you found the gap yourself — say so and leave it to a human to add, rather than building it here. Never open a PR or an issue in TauCetiRoadmap yourself; reviewing a roadmap change needs human attention.
Improving code that already exists is always in scope and needs no roadmap entry: refactoring, simplifying proofs, fixing or modestly generalising an existing lemma (without expanding the project's mathematical scope), relocating a misplaced declaration, adopting a cleaner idiom, and documenting existing code are all welcome at any time. The roadmap conditions what new mathematics gets added, not whether already-merged code may be made better.
Every pull request description must also contain one standalone attribution line:
Roadmap: CanonicalAreaName
or:
Roadmap: none
Use the canonical roadmap directory name. This is attribution, not authorization: new
mathematics must still cite the exact roadmap file and target it advances, while a refactor
needs no fresh roadmap claim but should name the one roadmap chiefly motivating it. Use
Roadmap: none for genuinely general, cross-cutting, infrastructure, or dependency work.
Do not infer the association from a TauCeti/ directory name; code organization and roadmap
scope do not coincide.
The rules of the repo
mainis always green. CI builds against pinned Mathlib and enforces: nosorry, no axioms beyondpropext,Classical.choice,Quot.sound(so nonative_decide), and the Mathlib linter set (style, file length, nomaxHeartbeatsoverrides). Do not try to disable these.- One topic per PR. Ship a prerequisite refactor as its own PR.
- Tau Ceti does not preserve backwards compatibility. When declarations or modules are moved,
renamed, replaced, or deleted — including when Mathlib supersedes them — update every
in-repository use and remove the obsolete names and module paths in the same PR. Do not add or
retain anything whose only purpose is compatibility: aliases, wrapper declarations, forwarding
import modules, deprecated shims (including
deprecated_module), or duplicate theorem names. External users of an older revision must update to the canonical API on currentmain; breaking their source compatibility is not a reason to keep an obsolete surface. TauCeti/is the only place code goes.scripts/,.github/, and the lakefile (lakefile.toml/lakefile.lean) are human-owned. The two Lake pins —lake-manifest.jsonandlean-toolchain— are an exception: a forward-only bump of them (Mathlib moving forward on the branch the lakefile nominates, with the toolchain moving monotonically forward) is machine-validated by thebump-guardcheck and is welcome, but never edit the lakefile or move a pin backward. Mathlib's lakefilerevmust always remainmaster;lake-manifest.jsonalone records the exact commit used by ordinary bumps and first-known-bad repair PRs.- Never delete a PR's human-owned changes to get it past the build gate. When a PR
intentionally touches
scripts/,.github/, or the lakefile (for example, a PR that adds a new CI check), those changes are the deliverable, not an obstacle. The gate routes such a PR to a human on purpose; the correct response is to wait for that human review and merge, never to strip the human-owned files so the PR looks auto-mergeable — that throws away the work the PR exists to do. This binds automated fix/review agents too: if a PR carries human-owned changes, leave it alone (skip it) rather than "fixing" it toward auto-merge. - Do not
--admin-merge AI-authored PRs. Landing a PR is the review pipeline's job: it merges only once every rubric is green (and, forTauCeti/-only diffs, CI is green). Using an admin override to bypass that gate — even when the reviews have not been run — defeats the project's quality control. If the pipeline is not producing verdicts, run the review (tauceti-review) or leave the PR for a human; never force it through.
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.
- 5d ago First seen · 99 lines · 1,483 tokens per session scan A 283c0cbc347a
TauCeti AGENTS.md is an instructions file published in the GitHub repository TauCetiProject/TauCeti (183 stars, last pushed today), licensed Apache-2.0. It adds 1,483 tokens to every session, about $0.0074 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
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
deepseek-harness AGENTS.md
AGENTS.md instructions for deepseek-ai/deepseek-harness, covering agents.md, pre-stable apis and released session data, repository layout, commands and host sandbox failures.