What is M87-Spine-lite CLAUDE.md?

A set of repository rules for operating as a restricted executor in the M87-Spine-lite project. It requires checks, receipts, quality gates, and audit verification for commands and file changes.

Use it when working in that repository to initialize sessions, check commands and file writes, record results, run quality checks, and close the session safely.

It prevents unapproved commands or writes and makes each session's actions traceable.

M87-Spine-lite CLAUDE.md — Instructions file for coding agents by MacFall7: A set of repository rules for operating as a…

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/macfall7/m87-spine-lite/claude-md

Clone the repo

git clone --depth 1 https://github.com/MacFall7/M87-Spine-lite

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/macfall7/m87-spine-lite/claude-md.svg)](https://agentmods.dev/instructions/macfall7/m87-spine-lite/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/macfall7/m87-spine-lite/claude-md"><img src="https://agentmods.dev/badge/instructions/macfall7/m87-spine-lite/claude-md.svg" alt="Measured on agentmods" height="20"></a>

Per session 797 This file is loaded in full into every session.

When invoked 797 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.

Model	Per session	Once invoked
Fable 5.1	$0.00797	$0.00797
Opus 5	$0.00398	$0.00398
Sonnet 5	$0.00159	$0.00159
Haiku 4.5	$0.00080	$0.00080

Measured 5d ago against content hash 82e22baa6388, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

M87-Spine-lite 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 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.

Makes network callslowCapability

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

**Network command reclassification:** Non-force `git push` is classified as `SHELL_MUTATING` (allowed under NORMAL posture). Force-push variants (`--force`, `-f`) are caught by `DENY_COMMANDS` and classified `SHELL_DANGE

CLAUDE.md · 68 lines

How it starts

The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md — Spine Lite Governance

This repo is governed. You operate as a bounded executor — proposal generator with scoped execution privileges, never the authority.

Non-negotiables

Fail-closed by default. Unknown → halt.
No network egress.
No writes outside approved scopes.
Every action produces a receipt (allowed or blocked).
Session closes only after quality gates + chain verification pass.

Required workflow (every session)

Initialize session: python hooks/governor.py init-session
Before ANY file write: python hooks/governor.py check-write --path <path> If DENY → do not write. Report the reason.
Before ANY shell command: python hooks/governor.py check-command --command "<command>" If DENY → do not execute. Report the reason.
After EACH action, emit receipt: python hooks/governor.py receipt --action <file_write|command|other> --path <path> --command "<cmd>" --exit-code <int>
Before modify/commit — quality gate: python hooks/governor.py quality-gate pre-modify python hooks/governor.py quality-gate pre-commit
Close session (verifies chain + audit): python hooks/governor.py close-session

Scope boundaries

Writable: src/, tests/, docs/, governance/, hooks/, schemas/, scripts/, config/, .github/workflows/

Denied: .env*, *.key, *.pem, *.secret, credentials*, node_modules/, .git/objects/, .git/refs/

Restricted (operator override required): governance/policy*.yaml, trust_zones.json, session_risk*, *.schema.json

Computer-use & MCP tool governance

Computer-use tools (computer screenshot, computer click, computer type, computer scroll, computer key, str_replace_based_edit_tool) are not in any allowlist. They fail-closed to SHELL_DANGEROUS and are always denied.

MCP browser tools (mcp__Claude_in_Chrome__*) follow the same path: unknown command → SHELL_DANGEROUS → DENY.

Network command reclassification: Non-force git push is classified as SHELL_MUTATING (allowed under NORMAL posture). Force-push variants (--force, -f) are caught by DENY_COMMANDS and classified SHELL_DANGEROUS before reaching any allowlist. All other network commands (curl, wget, ssh, git pull, git clone, pip install, npm install) remain NETWORK_ATTEMPT → DENY.

Read the full file on GitHub · 68 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.

5d ago First seen · 68 lines · 797 tokens per session scan A 82e22baa6388

Subscribe to this mod's changes

M87-Spine-lite CLAUDE.md is an instructions file published in the GitHub repository MacFall7/M87-Spine-lite (10 stars, last pushed 1mo ago), licensed MIT. It adds 797 tokens to every session, about $0.0040 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.

agent-governance ai-agents auditability claude-code cryptographic-receipts deterministic-execution

M87-Spine-lite CLAUDE.md

Getting it into your agent

Wrote this? Show the measurements

What it costs to keep this loaded

Grade A, and why

Makes network callslowCapability

How it starts

CLAUDE.md — Spine Lite Governance

Non-negotiables

Required workflow (every session)

Scope boundaries

Computer-use & MCP tool governance

What this file has done since we first saw it

Other instructions, from other repositories

FCoP AGENTS.md

FCoP CLAUDE.md

Meta_Kim AGENTS.md

Meta_Kim CLAUDE.md

DashClaw copilot-instructions.md

sofagent GEMINI.md