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/zerosixty/samurai/claude-mdgit clone --depth 1 https://github.com/zerosixty/samuraiWhat 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 | $0.03483 | $0.03483 |
| Opus 5 | $0.01741 | $0.01741 |
| Sonnet 5 | $0.00697 | $0.00697 |
| Haiku 4.5 | $0.00348 | $0.00348 |
Grade A, and why
samurai 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 — 296 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file contains guidance for Claude Code when working on this project.
Project Overview
Samurai is a scoped testing framework for Go with path isolation, inspired by GoConvey but modernized with:
- Explicit context passing (no GLS/goroutine-local-storage)
- Standard context.Context support for cancellation/timeouts
- Assertion-library agnostic (users bring their own)
- Zero external dependencies
- Parallel execution by default — including the top-level
TestXxxfunction, not just samurai-generated sub-tests. Passsamurai.Sequential()to opt out (required for tests that uset.Setenv, bind fixed ports, etc.) - No double execution - code runs exactly once per test path
- Generic
RunWithfor custom test contexts via embedded*BaseContext(Go 1.25+ generic type aliases)
Git Workflow
- NEVER push directly to master. Always work through a PR: create a branch, push,
gh pr create,gh pr merge. - Release via the
ReleaseGitHub Actions workflow (workflow_dispatch). Trigger it withgh workflow run release.yml -f version=vX.Y.Z(version must matchv*.*.*). The workflow gates the release ongo vet+go test -race ./..., then creates the tag and the GitHub Release with--generate-notes. NEVER tag or release manually (git tag,gh release create) — that bypasses the race-test gate. See.github/workflows/release.yml.
Build & Test Commands
go test ./... # Run all tests
go test -v ./... # Verbose test output
go test -race ./... # Run with race detector
Code Style
- Follow standard Go conventions
- Keep the API minimal - users control behavior via their assertion library and
t.Skip() - Prefer composition over inheritance
- All public types and functions must have doc comments
Architecture
samurai.go- Public API:Run,RunWith[V],Contextinterface,Wtype alias, options (Parallel,Sequential); internal path tree building (pathNode,buildPathTree,executeTree[V]);identityFactoryfor the non-genericRunscope.go-TestScope[V]generic struct withTest()andSkip()methods;type Scope = TestScope[W]aliasscope_runner.go- Generic path discovery (collectScopedPaths[V],collectPathsFromChildren,probeBuilder) and execution (executeScope[V],executeScopedPath[V])base_context.go-BaseContextstruct implementingContextinterface; providesContext(),Testing()andCleanup()runner.go- Shared helpers: panic recovery, cleanup execution
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 · 296 lines · 3,483 tokens per session scan A bbc307c47714
samurai CLAUDE.md is an instructions file published in the GitHub repository zerosixty/samurai (11 stars, last pushed 3mo ago), licensed MIT. It adds 3,483 tokens to every session, about $0.0174 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-go AGENTS.md
AGENTS.md instructions for avelino/awesome-go, covering awesome-go · llm contribution guide, project snapshot, when modifying the awesome list, coding guidelines and testing & validation.
ollama CLAUDE.md
Claude Code instructions for ollama/ollama: See AGENTS.md for the shared agent instructions for this repository.
azure-sdk-for-go go-code.instructions.md
Instructions for Azure/azure-sdk-for-go: All code should follow the guidelines from the Azure Go SDK Guidelines. This document is a summary of the most important guidelines to follow when contributing to the Azure Go SDK.
neva AGENTS.md
AGENTS.md instructions for nevalang/neva, covering neva engineering guide, project model, repository map, documentation and ai engineering harness.
chatgpt-cli CLAUDE.md
Instructions for kardolus/chatgpt-cli, covering chatgpt-cli — release runbook, prerequisites, 1. cut the release, 2. publish the github release + binaries and 3. update the homebrew tap.
homebutler CLAUDE.md
Instructions for Higangssh/homebutler, covering homebutler, verify before claiming anything works, writing commits, prs, and comments, no tool attribution and commit subjects.