cli-collaboration CLAUDE.md

A collaboration guide for coordinating several coding agents working on the same repository. It defines handoffs, file ownership, reserved areas, review rules, and checks before editing.

In plain words
What is it for?
Use it to start a shared task safely, read or update the handoff record, declare files to edit, and document tests, concerns, and the next step.
Why use it?
It prevents agents from overwriting each other’s work or changing files reserved for another person or session.

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/spe1977/cli-collaboration/claude-md
Clone the repo
git clone --depth 1 https://github.com/Spe1977/cli-collaboration
Per session 771 This file is loaded in full into every session.
When invoked 771 The same file — it is already loaded in full.
Security scan A 0 findings. 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 $0.00771 $0.00771
Opus 5 $0.00385 $0.00385
Sonnet 5 $0.00154 $0.00154
Haiku 4.5 $0.00077 $0.00077

Measured 2d ago against content hash 395b5952f601, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cli-collaboration 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

CLAUDE.md · 80 lines

How it starts

The opening of the file, as written. The whole thing — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

This project uses the cli-collaboration skill to coordinate Codex, Claude Code, Gemini CLI, Grok Build, and single-agent multi-session work on a shared repository.

Activation

When AGENT_HANDOFF.md exists at the project root, read it before any other file. Apply the cli-collaboration skill: declare the start gate, respect ### agent-owned, ### user-reserved, and ### frozen sections, and update the handoff last with files changed, tests, open concerns, and the next concrete step.

If AGENT_HANDOFF.md does not exist and the request involves a dirty worktree, work left by another CLI agent, or resume-from-previous-session intent, create the handoff in bootstrap mode (see skills/cli-collaboration/references/handoff-template.md, "Bootstrap variant") before any code change.

Start Gate

Before editing any file, state explicitly:

  • Handoff read: path and last-updated timestamp.
  • Current task: one line.
  • Files I will touch: explicit list, not directories.
  • Expected red test: name, or "no test for this task" with the reason.
  • Reserved zones confirmed: user-reserved and frozen areas read.
  • Stop condition: task-complete | context-budget | blocker.

If you cannot fill one of those fields, read more or ask the user. Do not improvise.

Ownership

Never edit a file listed under ### user-reserved or ### frozen without explicit user approval. For ### agent-owned files, edit only those assigned to Claude. If the request requires touching a file owned by Codex, Gemini, or another agent, stop, name the file and owner, and ask for reassignment or deferment.

When in doubt, run:

skills/cli-collaboration/scripts/check-ownership.sh --agent Claude <file...>

Exit 0 = no conflict, 1 = ownership conflict, 2 = malformed handoff.

Destructive Actions

Never run or recommend without explicit user approval:

  • git reset --hard
  • git clean
  • git stash (unauthorized)
  • git restore (on files not authored in this shift)
  • git checkout -- (on files not authored in this shift)
  • lateral overwrite of files authored by another agent or the user

Read the full file on GitHub · 80 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 · 80 lines · 771 tokens per session scan A 395b5952f601

Subscribe to this mod's changes

cli-collaboration CLAUDE.md is an instructions file published in the GitHub repository Spe1977/cli-collaboration (3 stars, last pushed 1mo ago), licensed MIT. It adds 771 tokens to every session, about $0.0039 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 2026-08-31.