agent-communication skill
✓ VerifiedEnables agents to discover, inspect context, and exchange messages with other active AI agents like Codex and Claude Code.
- GitHub stars
- 1,663
- Files
- 2
- Last commit
- Nov 9, 2026
- License
- Apache-2.0
- Allowed tools
- —
What does the agent-communication skill do?
agent-communication enables agents to discover, inspect context, and exchange messages with other active AI agents like Codex and Claude Code.
Summary generated from the skill's SKILL.md.
About the agent-communication skill
Developers building multi-agent workflows can use this skill to facilitate direct interaction between active AI agents like Codex and Claude Code. It allows an agent to identify other running processes, inspect their recent activity, and exchange specific data or requests. This is useful when one agent needs to coordinate with another to complete a complex task or share project context.
An agent should load this skill when it needs to find a peer using the list command, examine the last few messages of a specific agent via the detail command, or transmit information using the send command. The tool provides structured data through JSON outputs and supports sending large amounts of text by piping command output directly into a message. It relies on the ai-devkit utility, which can be executed via npx if it is not present in the system path.
What an agent can do with it
- 1Inspecting agent contextA developer can instruct an agent to read the last twenty messages from another agent to understand its current state.
- 2Forwarding command outputYou can send the results of a terminal command directly to another agent using standard input.
Written from the skill's SKILL.md. Every use case is checked against a quote from it.
How to install the agent-communication skill
Each command copies the skill folder into the client's skills directory. Restart the client or start a new session to load it.Install the agent-communication skill in Claude Code
~/.claude/skills/agent-communication/SKILL.md
$ npx degit codeaholicguy/ai-devkit/skills/agent-communication ~/.claude/skills/agent-communicationInstall the agent-communication skill in Claude Code (project)
.claude/skills/agent-communication/SKILL.md
$ npx degit codeaholicguy/ai-devkit/skills/agent-communication .claude/skills/agent-communicationInstall the agent-communication skill in Codex
~/.agents/skills/agent-communication/SKILL.md
$ npx degit codeaholicguy/ai-devkit/skills/agent-communication ~/.agents/skills/agent-communicationInstall the agent-communication skill in Gemini CLI
~/.gemini/skills/agent-communication/SKILL.md
$ npx degit codeaholicguy/ai-devkit/skills/agent-communication ~/.gemini/skills/agent-communicationInstall the agent-communication skill in Claude.ai
npx degit codeaholicguy/ai-devkit/skills/agent-communication agent-communication && zip -r agent-communication.zip agent-communication # claude.ai → Customize → Skills → + → Create skill → Upload a skill → agent-communication.zip
Automated checks
Ran Nov 15, 2026Full report →Passed our automated checks. That is not a security audit: read the scripts before you let an agent run them. How checks work
- Source repository linked and reachablecodeaholicguy/ai-devkit/skills/agent-communication is public and reachablegithub.com/codeaholicguy/ai-devkit/tree/HEAD/skills/agent-co ↗Pass
- Open source license (OSI)Apache-2.0spdx.org/licenses/Apache-2.0.html ↗Pass
- Commit or release in the last yearLast commit 2026-11-09github.com/codeaholicguy/ai-devkit/commits ↗Pass
- No download-and-run or credential theft in scriptsNo scripts in the skill folderPass
- No hidden instructions in SKILL.mdNo override phrases, safeguard switches, credential requests or hidden charactersgithub.com/codeaholicguy/ai-devkit/blob/HEAD/skills/agent-co ↗Pass
- Pre-approved tools are scopedNo pre-approved tools; the agent asks before using onegithub.com/codeaholicguy/ai-devkit/blob/HEAD/skills/agent-co ↗Pass
- SKILL.md follows the Agent Skills formatValid name and description; says when to use it; 27 linesgithub.com/codeaholicguy/ai-devkit/blob/HEAD/skills/agent-co ↗Pass
SKILL.md
View on GitHubAgent Communication
Use ai-devkit agent ... to discover and communicate with active agents. If ai-devkit is not on PATH, use npx ai-devkit@latest agent ....
Commands
ai-devkit agent list --json
ai-devkit agent detail --id <agent-name> --json --tail 20
ai-devkit agent send --id <agent-name> "<message>"
ai-devkit agent send --id <agent-name> --wait --timeout 120000 --json "<message>"
<command> 2>&1 | ai-devkit agent send --id <agent-name> --stdin
Notes
list --jsonreturns active agents with fields such asname,type,status,summary,projectPath, andlastActive.- Use the
namefromlist --jsonas--id. Partial matches are supported, but exact names are safer. - Use
detail --json --tail <n>to read recent context from an agent before deciding what to send. send --waitwaits for a reply; add--jsonwhen the response should be machine-readable.send --stdinforwards piped command output or larger text.
Files in this skill
2agent-communication skill questions
How do I install the agent-communication skill in Claude Code?
Run npx degit codeaholicguy/ai-devkit/skills/agent-communication ~/.claude/skills/agent-communication. Claude Code loads skills from ~/.claude/skills (all projects) or .claude/skills (one project) and uses a skill when its description matches the task.
Which clients can use the agent-communication skill?
Any client that reads the Agent Skills format: Claude Code, Codex, Gemini CLI and claude.ai (upload as a ZIP). Scripts in the skill run only where the client allows code execution.
How can I see a list of all currently active agents?
You can use the command ai-devkit agent list --json to retrieve a list of active agents including their name, type, and status.
How do I read the recent history or context of a specific agent?
Use the detail command with the agent's ID and a tail count to read recent context, for example: ai-devkit agent detail --id <agent-name> --json --tail 20.
Can I send the output of a command directly to another agent?
Yes, you can pipe command output to an agent using the --stdin flag with the send command.