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/block/trailblaze/claude-mdgit clone --depth 1 https://github.com/block/trailblazeWhat 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 | $0.00441 | $0.00441 |
| Opus 5 | $0.00220 | $0.00220 |
| Sonnet 5 | $0.00088 | $0.00088 |
| Haiku 4.5 | $0.00044 | $0.00044 |
Grade A, and why
trailblaze 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.
What it actually says
Working with AI Agents in this Repository
This file captures conventions for AI coding assistants (Claude Code, Cursor, Aider, etc.) working in this repo. Humans are welcome to read it too.
Commit attribution
Trailblaze attributes commits to the human developer who owns the change,
not to the AI assistant that helped author it. Please do not add
Co-Authored-By: trailers naming Claude, Cursor, GPT, or any other AI agent
to commit messages. The developer's name on the Author: line is the record
of authorship the project relies on for code review, blame, and release notes.
If you use an AI assistant while preparing a change, that's expected and
encouraged — just keep the assistant out of the commit metadata. The same
applies to PR descriptions: prose generated with AI help is fine, but don't
sign them on the assistant's behalf or include 🤖 Generated with … footers.
If you would like to acknowledge AI help on a particular change, do it in the PR description body in the developer's own words rather than as a structured trailer that downstream tooling will treat as a real author.
CLI invocations in docs
Public-facing docs (docs/**/*.md, examples/*/README.md) refer to the CLI as
trailblaze — the binary an installed user has on their PATH. This is
the audience the docs are written for.
Framework developers working out of a source checkout of this repo should
mentally swap trailblaze → ./trailblaze (the wrapper script at the OSS repo
root that rebuilds + dispatches to the source-built CLI). Don't bake ./ into
docs or example READMEs — that's a dev-loop detail, not a contract for the
audience we're writing for.
The same convention applies to commands an AI assistant suggests to the
developer in this repo: if you're walking a contributor through driving a
device from this checkout, use ./trailblaze. If you're updating user-facing
docs, use trailblaze.
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 · 39 lines · 441 tokens per session scan A 20093cda647c
trailblaze CLAUDE.md is an instructions file published in the GitHub repository block/trailblaze (305 stars, last pushed 7d ago), licensed Apache-2.0. It adds 441 tokens to every session, about $0.0022 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
daisyui code_generation_with_git.instructions.md
Instructions for saadeghi/daisyui, covering activation, initial setup, development loop, error handling and session end.
planning-with-files AGENTS.md
AGENTS.md instructions for OthmanAdi/planning-with-files, covering agents.md — planning-with-files agent reference card, commit rules, release checklist (12 steps), version bump scope and changelog format.
mq commit.instructions.md
Instructions for harehare/mq: Use the following format for commit messages.
stiletto CLAUDE.md
Instructions for benjamin-bader/stiletto, covering stiletto — contributor & agent guide, commit messages: conventional commits (required), breaking changes, examples and how releases work.
crow CLAUDE.md
Instructions for kh0pper/crow, covering claude.md, working with this repo, network exposure invariant, maintaining claude.md vs crow.md and where to find things.
cotect CLAUDE.md
Instructions for cotect-dev/cotect, covering project conventions and git commits.