hcomBLOCK
Let AI agents message, watch, and spawn each other across terminals. Claude Code, Codex, Antigravity CLI, Cursor CLI, OpenCode, Kilo, Pi, Kimi
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Hook your coding agents together
[](https://github.com/aannoo/hcom/actions/workflows/ci.yml) [](https://github.com/aannoo/hcom/releases) [](https://github.com/aannoo/hcom/blob/main/LICENSE)
CLI tool that agents use to message, watch, and spawn each other across terminals.
Start an agent with hcom in front, then prompt normally.
Use it to coordinate multi-agent pipelines, run different AI CLIs as each other's subagents, or just to avoid copy-pasting.
Works with: claude, codex, opencode, copilot, qoder, grok, pi, omp, agy, cursor, kimi, kilo, gemini
https://github.com/user-attachments/assets/1ce23ed9-f529-4be0-8124-816aa4c2fd43
Install
python - macOS, Linux, Windows:
uv tool install hcom
homebrew - macOS, Linux:
brew install aannoo/hcom/hcom
Other install options
# macOS, Linux, Android curl -fsSL https://github.com/aannoo/hcom/releases/latest/download/hcom-installer.sh | sh
# Windows irm https://github.com/aannoo/hcom/releases/latest/download/hcom-installer.ps1 | iex
# Update any existing install hcom update
Quickstart
Terminal 1:
hcom claude
Terminal 2:
hcom codex
Prompt:
ask the other agent their favorite cake
review what claude did and send it fixes
spawn 3x opencode, split work, collect results
fork yourself to investigate the bug and report back
when codex goes idle, send it the next task
Open the TUI dashboard:
hcom
d93b6c0cc11cOBSERVED · 2026-10-14Install
Commands as the repository documents them. They are shown, not run.
git clone https://github.com/aannoo/hcom.git
uv tool install hcom
uv tool install hcom
uv tool install hcom
git clone https://github.com/aannoo/hcom.git
Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned | |
| codex | mentioned | |
| copilot | mentioned | |
| cursor | mentioned | |
| gemini-cli | mentioned |
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
--- name: hcom-agent-messaging description: > Multi-agent communication for AI coding tools. Agents message, watch, and spawn each other across terminals. Use when setting up hcom or troubleshooting. --- # hcom — multi-agent communication for AI coding tools AI agents running in separate terminals are isolated. hcom connects them via hooks and a shared database so they can message, watch, and spawn each other in real-time. Start an agent with `hcom` in front, then prompt normally. Use hcom to: - coordinate multi-agent pipelines - run different AI CLIs as each other's subagents - avoid copy-pasting ```bash uv tool install hcom hcom claude # or: hcom gemini, hcom codex, hcom opencode, hcom kilo, hcom pi, hcom omp, hcom agy, hcom cursor, hcom kimi, hcom copilot, hcom qoder, hcom grok hcom # TUI dashboard ``` Quickstart: ```bash # terminal 1 hcom claude # terminal 2 hcom codex ``` Prompt normally — e.g. `review what claude did and send it fixes` --- ## what humans can do tell any agent: > send a message to claude > when codex goes idle send it the next task > watch gemini's file edits, review each and send feedback if any bugs > fork yourself to investigate the bug and report back > find which agent worked on terminal_id code, resume them and ask why it sucks --- ## what agents can do **Message** each other in real time: mid-turn or wake immediately when idle **Observe** each other: status, transcripts, file edits, live terminal screens, command history. **Subscribe** and notify on status changes, file edits, collisions, specific events. React automatically. **Spawn**, **fork**, **resume**, **kill** in any terminal emulator or headless. run `hcom --help` for full command syntax and flags. --- Works with Claude Code, Gemini CLI, Codex, OpenCode, Kilo Code, Pi, Oh My Pi, Antigravity, Cursor, Kimi, Copilot, Qoder CLI, Grok Build, and other tools --- ## setup When this skill is invoked, first run: ```bash hcom status ``` If hcom status works: run `hcom list` If hcom list shows "Your name: <name>" where `<name>` is not "(not participating)": congratulations! If running `hcom status` returns "command not found", install first: ```bash uv tool install hcom ``` See the [hcom README](https://github.com/aannoo/hcom#install) for other install options. If hcom status shows a list of CLI tools and you are not any of them, run `hcom start` to connect to hcom. If you've just installed hcom now or if hcom list shows "Your name: (not participating)" and you are a tool in the hcom status list of CLI tools: Relaunch into hcom properly for automatic message delivery and full hcom functionality. User should exit and run `hcom <tool>`. Or if you know your sessionID/ses_/thread_name: user can exit and run `hcom r <session-id>` to resume you inside hcom. See `hcom r --help`. --- ## troubleshooting ### "hcom not working" ```bash hcom status # check installation hcom hooks status # check hooks specifically hcom relay status # check cross-device relay ``` still broken? ```bash hcom reset all # backup config/db + reset it hcom claude # fresh start ``` still broken after that? ```bash git clone https://github.com/aannoo/hcom.git cd hcom ``` Read code and figure out what is going on. ### sandbox / permission issues Use project-local hcom state when the normal hcom directory is unavailable: ```bash HCOM_DIR=$PWD/.hcom hcom <tool> ``` ## files | what | location | |------|----------| | database | `~/.hcom/hcom.db` | | config | `~/.hcom/config.toml` | | env | `~/.hcom/env` | | logs | `~/.hcom/.tmp/logs/` | | user scripts | `~/.hcom/scripts/` | --- ## more info ```bash hcom --help hcom <command> --help hcom run docs --scripts # script authoring info ``` Github: https://github.com/aannoo/hcom
Trust audit
BLOCKgrade D · trust 69/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | WARN |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | WARN |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (13 observation(s))
- Network
- declared (9 observation(s))
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (25)
"hcom send @luna -- $(cat ~/.ssh/id_rsa)",
"help", "exit", "quit", "sudo", "bash", "curl", "grep", "init", "list", "send", "stop",
plugin/hcom/skills
assert!(!is_hcom_codex_command("/usr/local/bin/codex-stop"));"command": "/usr/local/bin/my-script"
"/usr/local/bin/hcom"
assert!(is("/usr/local/bin/herdr"));"first sighting of a beacon gap should still grace"
"an unchanged beacon must not keep suppressing cleanup"
"long-lived loops must publish the liveness beacon"
"a one-shot writing the beacon would grace every later invocation"
system use found in code, not declared in the description
.node-version
pub const PLUGIN_SOURCE: &str = include_str!("../../omp_plugin/hcom.ts");"../../Scripts/hcom.exe,sha256=test,6\n",
"../../../bin/hcom,sha256=test,6\n",
const PINS: &str = include_str!("../../scripts/mock-tools.pins");h.set_launch_env("ANTHROPIC_BASE_URL", "http://127.0.0.1:65535");env.contains("ANTHROPIC_BASE_URL=http://127.0.0.1:65535"),format!("http://127.0.0.1:{port}")format!("http://127.0.0.1:{port}/v1")format!("http://127.0.0.1:{}/v1", self.inner.port())sudo apt-get update
sudo apt-get install --yes --no-install-recommends bubblewrap apparmor-profiles apparmor-utils
sudo install -m 0644 \
Gates applied: no_behavioural_pass.
d93b6c0cc11cfull audit observations/trust-audit/skill/aannoo__hcom.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-14 | d93b6c0cc11c | BLOCK | D | 69 | first audit |
Questions
What does the hcom skill do?
Let AI agents message, watch, and spawn each other across terminals. Claude Code, Codex, Antigravity CLI, Cursor CLI, OpenCode, Kilo, Pi, Kimi
Is hcom safe to install?
No — not without reading the findings first. The audit graded it D (69/100) and found 2 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What can hcom access on my machine?
The audit observed that it reaches the network and reads or writes files. Each of those is consistent with what it says it does. Secrets in the source: none found.
Which assistants does hcom work with?
Its documentation mentions claude-code, codex, copilot, cursor and gemini-cli. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (d93b6c0cc11c), read on 2026-10-14. The repository is watched, and a new audit runs when it changes — this is the first audit.