mind CLAUDE.md

Repository instructions for a tool that manages coding-agent add-ons and checks that the written specification matches the tests and implementation.

In plain words
What is it for?
Use them when adding or changing features in this Rust project, especially when updating the specification, tests, agent homes, or installed tools.
Why use it?
They keep feature behavior documented and tested together, while catching undocumented or untested specification rules.

Instructions file

Install

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.

agentmods
npx agentmods add instructions/jaemk/mind/claude-md
Clone the repo
git clone --depth 1 https://github.com/jaemk/mind
Per session 4,409 This file is loaded in full into every session.
When invoked 4,409 The same file — it is already loaded in full.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.04409 $0.04409
Opus 5 $0.02204 $0.02204
Sonnet 5 $0.00882 $0.00882
Haiku 4.5 $0.00441 $0.00441

Measured 2d ago against content hash 35dd407f6ce8, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mind CLAUDE.md scanned grade A with 1 finding 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

`PATH` (see the fake `curl`/`gh` tests in `src/selfupdate.rs`).
CLAUDE.md · 279 lines

How it starts

The opening of the file, as written. The whole thing — 279 lines — stays where its author put it; the contents beside it link to each section on GitHub.

mind

A manager for agent tooling (skills, agents, rules, tools) that melds with arbitrary git repos and links installed items into one or more agent homes (default ~/.claude; see Paths::agent_homes). A tool is store-only: it is referenced by other items, not linked into an agent home.

The behavioral spec lives in spec/ and is the reference to verify against. Each statement there has a stable ID (e.g. LIFE-4); tests cite the IDs they cover in // spec: ID comments. Keep the spec and the implementation in step: when behavior changes, update the spec doc and the citing tests in the same change.

tests/spec_coverage.rs is a coverage gate run by cargo test (and CI): it fails when a defined spec ID is neither cited by a test nor listed in its ALLOWLIST, and when a test cites an ID that the spec does not define. So a new spec ID forces a coverage decision - add a citing test or an allowlist entry with a reason - and a test cannot cite an undocumented behavior. Only citations in test code count: everything under tests/, plus the #[cfg(test)] regions of src/. A // spec: comment on production code documents the implementation site and is fine to write, but it asserts nothing, so it does not satisfy the gate. CI is .github/workflows/ci.yml (runs make ci: fmt-check + clippy + test). Locally, use make ci-local (alias: make check): it runs the same clippy + test gate but formats in place (cargo fmt) instead of fmt-check, so one command both fixes formatting and runs the full gate. Prefer make ci-local over chaining cargo fmt and make ci.

The toolchain is pinned in rust-toolchain.toml, so make ci-local and make ci lint with the same clippy. Do not bump it casually: a newer clippy adds lints, -D warnings makes each one a hard error, and the new errors usually land on code the bump does not touch. To bump, change channel there, update the matching toolchain: inputs in .github/workflows/{ci,release,canary}.yml, and run make ci-local. The jobs that build against a different toolchain on purpose (the MSRV check, the release build, the crates.io publish) set RUSTUP_TOOLCHAIN, which outranks the file; leave that in place or those gates quietly start testing the wrong compiler.

Read the full file on GitHub · 279 lines

Changes

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.

  1. 2d ago First seen · 279 lines · 4,409 tokens per session scan A 35dd407f6ce8

Subscribe to this mod's changes

mind CLAUDE.md is an instructions file published in the GitHub repository jaemk/mind (1 stars, last pushed 3d ago), licensed MIT. It adds 4,409 tokens to every session, about $0.0220 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.