Skip to content
Agentmart
Submit a tool

fleet-send

Send a message to another machine in the fleet — context, task, alert, or broadcast with 7-priority fallback delivery

Skill by supportersimulator

IndexedMIT★ 2
Write a review

Chalk Labs

Need help deploying fleet-send?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/supportersimulator-fleet-send

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Send a message to another machine in the fleet — context, task, alert, or broadcast with 7-priority fallback delivery

Use cases: productivity

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/fleet-send

What's inside

Skill name
fleet-send
Description
Send a message to another machine in the fleet — context, task, alert, or broadcast with 7-priority fallback delivery
License
MIT
Instructions
224 lines in SKILL.md · view on GitHub

Before installing

  • Instructions call the network

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Fleet Send

Send a message to one or more machines in the fleet. Messages are routed by type — each type has different delivery semantics. The 7-priority fallback chain ensures delivery even on hostile networks.

## Protocol Awareness (INVARIANT)

**Always try channels in order. Never skip to P8.**

The send system follows the P0-P8 cascade strictly. Each channel gets a 30-second timeout before falling through to the next. The order is: P1 NATS, P2 HTTP, P3 Chief relay, P4 Seed file, P5 SSH, P6 WoL, P7 Git, P8 Direct text (last resort only).

**If your message used P3+, also trigger repair for the broken channels.** A P3+ delivery is a success AND a signal that P1/P2 need fixing. The daemon handles this automatically, but if you're sending manually, check channel state and invoke repair:

```bash
# Check channel state before/after sending
curl -s http://127.0.0.1:8855/channels | python3 -m json.tool

# If channels are broken, trigger repair
curl -sf -X POST http://127.0.0.1:8855/repair -H "Content-Type: application/json" \
  -d '{"target": "<node-id>", "channel": "nats"}'
```

## How It Works

1. You compose a message with a type, target, and payload
2. The daemon tries all channels in priority order (P1 NATS through P7 Git, 30s timeout each)
3. First successful delivery wins — the channel used is logged
4. If P1-P2 fail and a lower channel succeeds, a self-healing repair request is sent automatically
5. Target processes the message based on type (seed file, agent spawn, notification, etc.)
6. **P8 (direct text/copy-paste) is NEVER used by the daemon** — it exists only as a human last resort

## Sending Messages

### Direct peer-to-peer (preferred)

```bash
# Send context (passive — injected on next prompt)
curl -sf -X POST http://<target-ip>:8855/message -H "Content-Type: application/json" \
  -d '{"type":"context","from":"<this-node>","to":"<target>","payload":{"subject":"Auth bug fixed","body":"Fixed the session token race condition in middleware.py line 42. You can pull main now."}}'

# Send task (autonomous — spawns agent on target)
curl -sf -X POST http://<target-ip>:8855/message -H "Content-Type: application/json" \
  -d '{"type":"task","from":"<this-node>","to":"<target>","payload":{"subject":"Run 3s cross-exam on auth module","body":"Full cross-examination of authentication middleware. Check session token storage compliance. Report findings."}}'

# Send alert (interrupt — focuses VS Code + macOS notification)
curl -sf -X POST http://<target-ip>:8855/message -H "Content-Type: application/json" \
  -d '{"type":"alert","from":"<this-node>","to":"<target>","payload":{"subject":"Conflicting verdicts on branch X","body":"STOP: mac2 and mac3 produced contradictory test results on feature/auth. Need human adjudication."}}'

# Send reply (references an original message)
curl -sf -X POST http://<target-ip>:8855/message -H "Content-Type: application/json" \
  -d '{"type":"reply","from":"<this-node>","to":"<target>","payload":{"subject":"Re: cross-exam results","body":"3 issues found. See /tmp/atlas-agent-results/fleet-task-abc123.log","ref":"msg-abc123"}}'

# Broadcast to all nodes
curl -sf -X POST http://127.0.0.1:8855/message -H "Content-Type: application/json" \
  -d '{"type":"broadcast","from":"<this-node>","payload":{"subject":"Pull main","body":"New feature landed — fleet-worker tmux isolation. Everyone pull and restart daemons."}}'
```

### Via CLI script

```bash
# Basic send (uses chief relay as transport)
CHIEF_INGEST_URL=http://<chief-ip>:8844 ./scripts/fleet-send.sh <target> "subject" "body"

# With priority (p1=alert, p2=task, p3=context)
CHIEF_INGEST_URL=http://<chief-ip>:8844 ./scripts/fleet-send.sh <target> "subject" "body" p1

# Broadcast to all nodes
./scripts/fleet-send.sh all "Pull main" "New feature landed"
```

### Via Python

```python
import requests, json

def fleet_send(target: str, msg_type: str, subject: str, body: str, ref: str = None):
    """Send a fleet message to a target node."""
    payload = {"subject": subject, "body": body}
    if ref:
        payload["ref"] = ref

    msg = {
        "type": msg_type,
        "from": os.environ.get("MULTIFLEET_NODE_ID", "unknown"),
        "to": target,
        "payload": payload
    }

    # Try direct peer-to-peer first
    try:
        r = requests.post(f"http://{target_ip}:8855/message",
                         json=msg, timeout=3)
        if r.status_code == 200:
            return {"delivered": True, "channel": "http_direct"}
    except requests.ConnectionError:
        pass

    # Fallback to chief relay
    try:
        r = requests.post(f"http://{chief_ip}:8844/message",
                         json=msg, timeout=5)
        if r.status_code == 200:
            return {"delivered": True, "channel": "chief_relay"}
    except requests.ConnectionError:
        pass

    return {"delivered": False, "error": "all channels failed"}
```

## Message Types

| Type | Delivery | Disrupts? | When to use |
|------|----------|-----------|-------------|
| `context` | Seed file, injected on next prompt | No | Enriching another session's context |
| `task` | Spawns `claude -p` agent on target | No | Autonomous work on another machine |
| `reply` | Seed file with `ref` linking to original | No | Responding to a specific request |
| `alert` | Focus VS Code + macOS notification | Yes | Urgent, needs human attention now |
| `broadcast` | Seed file on ALL nodes | No | Fleet-wide announcements |
| `sync` | Silent state bookkeeping | No | Internal coordination (heartbeats, acks) |

From supportersimulator/multi-fleet, MIT.

More from supportersimulator/multi-fleet

Signals

Stars
2
Forks
0
Open issues
0
Last commit (this folder)
Apr 5, 2026 (7 months ago)
Last release
Apr 18, 2026 (6 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from supportersimulator

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get fleet-send running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying fleet-send?

CallGet help
Cite this page

fleet-send skill. Agent Mart. https://agentmart.dev/skills/supportersimulator-fleet-send (last updated 2026-10-24)