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/agentic-research/mache/claude-mdgit clone --depth 1 https://github.com/agentic-research/macheWrote 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/agentic-research/mache/claude-md)<a href="https://agentmods.dev/instructions/agentic-research/mache/claude-md"><img src="https://agentmods.dev/badge/instructions/agentic-research/mache/claude-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 | $0.01914 | $0.01914 |
| Opus 5 | $0.00957 | $0.00957 |
| Sonnet 5 | $0.00383 | $0.00383 |
| Haiku 4.5 | $0.00191 | $0.00191 |
Grade A, and why
mache 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 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 — 88 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
Build & Test Commands
All commands use Task.
task build # Build binary + codesign (macOS)
task test # Run all tests (go test -v ./...)
task lint # golangci-lint run ./...
task fmt # gofumpt -w -extra .
task vet # go vet ./...
task check # full CI-equivalent: fmt + vet + lint + parity + test + validate + docs/actions/yaml lints + smells + gates:preflight + server-json
task validate # SQLite ingestion tests only
task test-go-schema # Self-hosting smoke test (ingests mache's own source)
task tidy # go mod tidy
Run a single test: task test -- -run TestName or go test -v -run TestName ./graph/.
Integration tests require real SQLite databases and are gated behind env vars (MACHE_TEST_KEV_DB, MACHE_TEST_NVD_DB) — they skip automatically when unset.
Architecture
Mache projects structured data (JSON, SQLite, source code) as a filesystem driven by declarative JSON schemas.
Three Presentation Paths
.db files → SQLiteGraph (zero-copy, lazy scan, direct SQL) → GraphFS (NFS) or MCP
other files → Engine (walkers + loaders) → MemoryStore → GraphFS (NFS) or MCP
any source → Engine → MemoryStore/SQLiteGraph → MCP (Streamable HTTP or stdio)
The .db path (SQLiteGraph + internal/template) is pure Go — no CGO, no tree-sitter. The source code path (Engine + SitterWalker) requires CGO for tree-sitter grammars. When ley-line pre-parses source into .db files with _ast tables, ASTWalker replaces SitterWalker and the entire pipeline is pure Go.
The mache serve command exposes the graph as MCP tools without mounting a filesystem.
Two transports: Streamable HTTP (--http :PORT, default, always-on) and stdio (--stdio, subprocess mode).
HTTP mode uses stateful sessions — each client gets its own session ID mapped to the projected graph.
Prefer HTTP over stdio to share one daemon across all sessions (avoids per-client FD leaks).
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 · 88 lines · 1,914 tokens per session scan A 227d6d73be18
mache CLAUDE.md is an instructions file published in the GitHub repository agentic-research/mache (45 stars, last pushed 4d ago), licensed Apache-2.0. It adds 1,914 tokens to every session, about $0.0096 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
agent-lsp CLAUDE.md
Instructions for blackwell-systems/agent-lsp, a project described as: MCP server that orchestrates language servers into agent-native workflows. 65 tools, 30 CI-verified languages.
siyuan-cli CLAUDE.md
Instructions for cicbyte/siyuan-cli, covering claude.md, 项目概述, 构建与运行, 开发构建 and 直接运行.
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.
repobrain copilot-instructions.md
Copilot instructions for study8677/repobrain, covering github copilot bootstrap instructions and hard rule — query the repobrain hub first.
gorest AGENTS.md
Instructions for pilinux/gorest, covering agents.md, project overview, build and run commands, build and tidy dependencies.