Answers from the documentation; checking which model is on duty
Hi - I answer from the OpenSmartRoute documentation: routing, the API, plans and quotas, self-hosting. Ask away, or open a support ticket if you need a person.
Grounded in the docs - follow a source before acting on it.
fleet - Skill - OpenSmartRoute
Skillv1.0.0
fleet
Fleet orchestration for distributed coding agents across Azure VMs. Invoked as /fleet <command>. Covers all fleet operations: status, scout, advance, adopt, watch, snapshot, dry-run, start, add-task…
Imported from rysweet/amplihack (.claude/skills/fleet/SKILL.md). Install upstream with npx skills add rysweet/amplihack --skill fleet. Copyright stays with the author.
/fleet
Fleet orchestration — manage coding agents (Claude Code, Copilot, Amplifier) running across multiple Azure VMs via azlin.
Invoke as /fleet <command> or just describe what you want and Claude will pick the right command.
Sequential VM polling: Session discovery from azlin list (no SSH). Pane capture via sequential SSH
Cached SSH output: Scout caches Phase 1 tmux captures for Phase 3 reasoning (no double-poll)
Incremental scout: --incremental flag skips unchanged sessions using ~/.amplihack/fleet/last_scout.json
Bastion tunnel reuse: Reuses existing SSH tunnels via get_existing_tunnels() instead of creating new ones
PR URL detection: Uses gh pr list on remote VM for reliable PR detection from git state
Health metrics in reasoning: fleet_health.py wired into SessionContext for admiral decisions
Unified status classifier: Single canonical classifier in _status.py (no dual TUI/CLI divergence)
Modular CLI: Commands split across _cli_session_ops.py, _cli_scout_advance.py, _cli_formatters.py (each under 400 LOC)
Project grouping: Scout report groups sessions by registered project with open objectives displayed
How to Run
Execute via Bash:
fleet <command> [options]
Presenting Results
After running fleet scout or fleet advance, present the output to the user as:
Summary table — reformat the CLI output into a clean markdown table with columns:
VM | Session | Status | Action | Conf | Summary
Proposed inputs — for any send_input decisions, show what the admiral wants to send
Follow-up commands — always end with the actionable next steps from the report:
fleet advance to send next command to all sessions
fleet advance --confirm to review each before executing
fleet advance --session vm:session to advance one specific session
fleet watch <vm> <session> to inspect specific sessions
Installed into a catalogue, chosen by a router
Install fleet and it becomes one more candidate the router can pick - when it fits.
A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.
When a provider is down or over quota the request moves on; the trace says it did.
Use it
Copy one of these into your project. Installing also returns the manifest and these snippets.
# after Install: the listing is in your workspace's routing pool - a plan picks it for its slotcurl-s-X POST https://api.opensmartroute.ai/api/v1/route -H 'Authorization: Bearer $OSR_API_KEY' -H 'Content-Type: application/json' -d '{"text": "...", "plan": true}'
Manifest
An Open Capability Manifest: the router reads it to know what this does, what it costs and when to pick it.
rysweet-amplihack-fleet.ocm.jsonjson
{
"ocm": "1",
"id": "rysweet-amplihack-fleet",
"kind": "skill",
"name": "fleet",
"description": "Fleet orchestration for distributed coding agents across Azure VMs. Invoked as `/fleet <command>`. Covers all fleet operations: status, scout, advance, adopt, watch, snapshot, dry-run, start, add-task, queue, auth, dashboard, tui, and more. Use when: user mentions fleet, agents, VMs, sessions, or asks \"what are my agents doing\".",
"publisher": "rysweet",
"version": "1.0.0",
"capabilities": {
"domains": [
"data_analysis",
"coding"
],
"tags": [
"skill-md",
"skills-sh"
],
"languages": [
"en"
]
},
"quality_prior": 0.6,
"examples": [
"Fleet orchestration for distributed coding agents across Azure VMs. Invoked as `/fleet <command>`. Covers all fleet operations: status, scout, advance, adopt, watch, snapshot, dry-run, start, add-task, queue, auth, dashboard, tui, and more. Use when: user mentions fleet, agents, VMs, sessions, or asks \"what are my agents doing\"."
],
"primary": false,
"metadata": {
"source": {
"provider": "skills.sh",
"repository": "https://github.com/rysweet/amplihack",
"path": ".claude/skills/fleet/SKILL.md",
"ref": "HEAD",
"url": "https://github.com/rysweet/amplihack/blob/HEAD/.claude/skills/fleet/SKILL.md",
"key": "rysweet/amplihack/.claude/skills/fleet/SKILL.md"
}
},
"instructions": "# /fleet\n\nFleet orchestration — manage coding agents (Claude Code, Copilot, Amplifier) running across multiple Azure VMs via azlin.\n\nInvoke as `/fleet <command>` or just describe what you want and Claude will pick the right command.\n\n## All Commands\n\n### Status & Monitoring\n\n| Command | What it does |\n| ---------------------------- | --------------------------------------------- |\n| `fleet status` | Quick text overview of all VMs and sessions |\n| `fleet scout` | Discover, adopt, dry-run reason, show report |\n| `fleet snapshot` | Point-in-time capture of all managed sessions |\n| `fleet watch <vm> <session>` | Live snapshot of a single session |\n| `fleet observe <vm>` | Observe sessions with pattern classification |\n| `fleet tui` | Interactive Textual TUI dashboard |\n\n### Admiral Control\n\n| Command | What it does |\n| ---------------- | ------------------------------------------------- |\n| `fleet advance` | **LIVE** — reason and execute actions on sessions |\n| `fleet dry-run` | Show what the admiral would do (no action) |\n| `fleet run-once` | Single PERCEIVE->REASON->ACT cycle |\n| `fleet start` | Run autonomous admiral loop |\n\n### Session Management\n\n| Command | What it does |\n| ------------------ | --------------------------------------------- |\n| `fleet adopt <vm>` | Bring existing tmux sessions under management |\n| `fleet auth <vm>` | Propagate auth tokens (GitHub, Azure, Claude) |\n\n### Task Management\n\n| Command | What it does |\n| ------------------------- | ---------------------------- |\n| `fleet add-task \"prompt\"` | Queue work for the fleet |\n| `fleet queue` | Show task queue |\n| `fleet dashboard` | Project-level tracking |\n| `fleet report` | Generate fleet status report |\n\n### Projects & Knowledge\n\n| Command | What it does |\n| -------------------------------------- | ------------------------------------------- |\n| `fleet project add/list/remove` | Manage fleet projects |\n| `fleet project add-issue <proj> <num>` | Track a GitHub issue as a project objective |\n| `fleet project track-issue <proj>` | Sync objectives from GitHub issues by label |\n| `fleet graph` | Fleet knowledge graph summary |\n\n### Co-Pilot\n\n| Command | What it does |\n| ---------------------- | ----------------------------- |\n| `fleet copilot-status` | Show copilot lock/goal state |\n| `fleet copilot-log` | Show copilot decision history |\n\n## Quick Reference\n\n| User says | Command |\n| ---------------------------------- | ---------------------------------- |\n| \"What are my agents doing?\" | `fleet scout` |\n| \"Show me the fleet\" | `fleet status` |\n| \"Send next steps to all sessions\" | `fleet advance` |\n| \"Advance without confirmation\" | `fleet advance --force` |\n| \"Watch what dev/cybergym is doing\" | `fleet watch dev cybergym` |\n| \"Add auth to the new VM\" | `fleet auth <vm>` |\n| \"Queue this task for the fleet\" | `fleet add-task \"prompt\"` |\n| \"Track issue #42 for myapp\" | `fleet project add-issue myapp 42` |\n| \"Sync objectives from GitHub\" | `fleet project track-issue myapp` |\n| \"Open the dashboard\" | `fleet tui` |\n\n## Key Options\n\n```\nfleet scout [--session vm:session] [--vm VM] [--skip-adopt] [--incremental] [--save PATH]\nfleet advance [--session vm:session] [--vm VM] [--force] [--save PATH]\nfleet dry-run [--vm VM ...] [--backend auto|anthropic|copilot]\nfleet adopt <vm> [--sessions s1 s2]\nfleet watch <vm> <session> [--lines 30]\nfleet auth <vm> [--services github azure claude]\nfleet add-task \"prompt\" [--priority high] [--repo URL]\nfleet start [--interval 300] [--max-cycles 10]\n```\n\n## Environment\n\n| Variable | Required for |\n| ------------------- | --------------------------------------------- |\n| `AZLIN_PATH` | All commands (auto-detected if azlin on PATH) |\n| `ANTHROPIC_API_KEY` | scout, advance, dry-run, run-once, start |\n\n## Admiral Configuration\n\n- **Model**: Claude Opus 4.6 (`claude-opus-4-6`)\n- **Max output tokens**: 128,000 (reasoning JSON)\n- **Context gathered per session**: full tmux scrollback + first 50 & last 200 transcript messages\n- **Safety**: 57 dangerous-input patterns blocked (with safe allow-list), confidence thresholds (60% send_input, 80% restart), --confirm default on advance\n- **Docs**: `docs/fleet-orchestration/ADMIRAL_REASONING.md`\n\n## Status Icons\n\n| Icon | Status | Meaning |\n| ----- | ------------------ | ----------------------------------------- |\n| `[~]` | thinking | Agent is actively processing |\n| `[>]` | running | Agent producing output |\n| `[.]` | idle | Agent at prompt, waiting for direction |\n| `[X]` | shell (dead agent) | No agent detected in this session |\n| `[Z]` | suspended | Agent backgrounded but alive |\n| `[!]` | error | Error detected in session output |\n| `[+]` | completed | Agent finished its task |\n| `[?]` | waiting input | Agent asked a question, awaiting response |\n\n## Performance & Architecture\n\n- **Sequential VM polling**: Session discovery from `azlin list` (no SSH). Pane capture via sequential SSH\n- **Cached SSH output**: Scout caches Phase 1 tmux captures for Phase 3 reasoning (no double-poll)\n- **Incremental scout**: `--incremental` flag skips unchanged sessions using `~/.amplihack/fleet/last_scout.json`\n- **Bastion tunnel reuse**: Reuses existing SSH tunnels via `get_existing_tunnels()` instead of creating new ones\n- **PR URL detection**: Uses `gh pr list` on remote VM for reliable PR detection from git state\n- **Health metrics in reasoning**: `fleet_health.py` wired into `SessionContext` for admiral decisions\n- **Unified status classifier**: Single canonical classifier in `_status.py` (no dual TUI/CLI divergence)\n- **Modular CLI**: Commands split across `_cli_session_ops.py`, `_cli_scout_advance.py`, `_cli_formatters.py` (each under 400 LOC)\n- **Project grouping**: Scout report groups sessions by registered project with open objectives displayed\n\n## How to Run\n\nExecute via Bash:\n\n```bash\nfleet <command> [options]\n```\n\n## Presenting Results\n\nAfter running `fleet scout` or `fleet advance`, present the output to the user as:\n\n1. **Summary table** — reformat the CLI output into a clean markdown table with columns:\n VM | Session | Status | Action | Conf | Summary\n\n2. **Proposed inputs** — for any `send_input` decisions, show what the admiral wants to send\n\n3. **Follow-up commands** — always end with the actionable next steps from the report:\n - `fleet advance` to send next command to all sessions\n - `fleet advance --confirm` to review each before executing\n - `fleet advance --session vm:session` to advance one specific session\n - `fleet watch <vm> <session>` to inspect specific sessions",
"cost": {
"context_tokens": 1898
}
}
Fetch it by URL: GET /api/v1/registry/rysweet-amplihack-fleet/manifest?version=1.0.0
Reviews
Star ratings from people who tried it. One review per account; edit yours any time.
No reviews yet. Install it, try it, and be the first to rate it.