agentleFS
buddyhbuddyh/agent-skills/agent-comms/SKILL.md

agent-comms

Skill5 starsChanged 8 months agoMITView on GitHub

Send messages to other Claude Code or Codex sessions via tmux. Hand off complex debugging context, get second opinions, share detailed technical findings across sessions.

  • Installs packages

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. elijahmuraoka/skills★ 58same name
    2. komluk/scaffolding★ 16same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository9 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read their context: before sending:

    line 65
  2. Very specific:

    Always include the [AGENT-MSG] prefix

    line 136

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Agent Communication

Send messages to and read output from other tmux sessions running Claude Code or Codex. Built for coordination between parallel agent sessions.

Platform Support

Works identically on both platforms -- the underlying mechanism is tmux, not the AI agent. Sessions can communicate across platforms (a Claude Code session can message a Codex session and vice versa).

Setup

Make the script executable (one-time):

chmod +x ~/.claude/skills/agent-comms/scripts/agent-msg

Message Format

Always prefix messages with a header so the receiving session knows it's agent-to-agent:

[AGENT-MSG] From: <your-session-name>
<your actual message here>

Get your session name:

tmux display-message -p '#S'

Commands

SCRIPT=~/.claude/skills/agent-comms/scripts/agent-msg

# List available sessions
$SCRIPT list

# Send message (types into their terminal + Enter)
$SCRIPT send <session> "<message>"

# Read their terminal output
$SCRIPT read <session> --lines 50

# Get their working directory
$SCRIPT info <session>

Workflow

  1. Get your session name for identification:
    tmux display-message -p '#S'
  2. List sessions to find the target:
    ~/.claude/skills/agent-comms/scripts/agent-msg list
  3. Read their context before sending:
    ~/.claude/skills/agent-comms/scripts/agent-msg read worker-1 --lines 50
  4. Send your message with identification:
    ~/.claude/skills/agent-comms/scripts/agent-msg send worker-1 "[AGENT-MSG] From: main-session
    
    Hit a wall on the websocket reconnection logic. Here's where I'm stuck: [detailed context...]"
  5. Read their response:
    ~/.claude/skills/agent-comms/scripts/agent-msg read worker-1 --lines 30

Example: Complex Debug Handoff

You've been debugging a native module issue for 20 minutes. It's a multi-layered problem involving Node versions, native rebuilds, and cross-project dependencies. You want another agent to take a look with full context:

# Check what sessions are running
~/.claude/skills/agent-comms/scripts/agent-msg list
# SESSION                  WINDOWS  ATTACHED
# worker-debug             1        no
# main-session             1        yes

# Send the full technical context
~/.claude/skills/agent-comms/scripts/agent-msg send worker-debug "[AGENT-MSG] From: main-session

Need a second opinion on a native module conflict. Here's what I've traced:

THE ISSUE:
- electron-builder fails with 'better-sqlite3 was compiled against a different Node.js version'
- This started after I ran nvm use 22 in another terminal for a different project

WHAT I'VE FOUND:
1. Main project expects Node 20.x for native modules
2. better-sqlite3 got rebuilt against Node 22 when I ran npm install in the other project
3. They share a global .node-gyp cache at ~/.node-gyp
4. The rebuild pulled in headers for 22.x, now electron can't load the module

WHAT I'VE TRIED:
- npm rebuild better-sqlite3 (still uses cached 22.x headers)
- Deleted node_modules, fresh install (same issue)
- electron-rebuild --force (fails with same ABI mismatch)

QUESTION:
Do I need to clear the global .node-gyp cache? Or is there a way to force better-sqlite3 to rebuild with electron's Node headers specifically? Wondering if I'm missing something obvious."

# Wait for them to analyze, then check response
~/.claude/skills/agent-comms/scripts/agent-msg read worker-debug --lines 50

When to Use This

  • Complex debugging: You've built up 15 minutes of context tracing an issue - share it rather than have another agent start from scratch
  • Cross-project conflicts: Version mismatches, shared caches, native module issues that span multiple repos
  • Architectural decisions: "Here's the tradeoff I'm seeing between approach A and B, with all the constraints I've discovered"
  • Rubber duck with context: Sometimes you just need another agent to see your full reasoning and spot the flaw

Why the Header Matters

The receiving session sees your message as user input. Without the [AGENT-MSG] header, it might think a human is typing. The header tells it:

  • This is automated agent-to-agent communication
  • Which session sent it (for replies)

Notes

  • Session names follow patterns like terminal-HHMMSS or custom names you assign
  • Always include the [AGENT-MSG] prefix
  • Be specific about what you need
  • The other agent can reply using the same skill

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.