Run a Dawn strategy lifecycle

Goal

Execute a complete dawn strategy workflow: install/check CLI, authenticate, prepare funding, create and iterate strategy code, launch paper/live runs, monitor status, and stop safely when requested.

When to use

Use this skill when the user asks to: - create a strategy from plain-English intent, - revise or upload strategy code, - launch paper or live strategy runs, - monitor run health/positions/logs, - stop or debug an active run.

Install and preflight

Install dawn if needed:

npm install -g @dawnai/cli

Verify:

# Print current Dawn CLI version
dawn version
dawn --help

Local source workflow only:

cd dawn-cli
npm install
npm run build
./install.sh

Command map

Auth: - dawn auth login - dawn auth status - dawn auth logout

Account: - dawn account overview - dawn account fund - dawn account wallet

Strategy authoring: - dawn strategy list - dawn strategy create "<text>" - dawn strategy status <conversationId> - dawn strategy revise <conversationId> "<text>" - dawn strategy rules <conversationId> list - dawn strategy rules <conversationId> approve <rule-index> - dawn strategy rules <conversationId> approve-all - dawn strategy code <conversationId> status - dawn strategy code <conversationId> generate - dawn strategy code <conversationId> export [--out <path>] [--json] - dawn strategy code <conversationId> upload <path-to-file>

Launch and operations: - dawn strategy launch <conversationId> --budget <usd> [--live] [--hours N] - dawn strategy positions <conversationId> [--strategy-id <strategyId>] - dawn run list - dawn run status <conversationId> - dawn run logs <conversationId> [--limit N] - dawn run stop <conversationId>

Standard flow

1. Authenticate: dawn auth login.
2. Confirm funding path: dawn account fund (required for live runs).
3. Create strategy: dawn strategy create "<request>" and capture conversationId.
4. Iterate strategy:
5. revise prompt (strategy revise) and/or upload files (strategy code ... upload),
6. review/approve rules,
7. generate code,
8. export code when needed (--json for multi-file map).
9. Launch:
10. paper: dawn strategy launch <conversationId> --budget 50
11. live: dawn strategy launch <conversationId> --budget 50 --live
12. custom duration: add --hours N
13. Monitor:
14. dawn run status <conversationId>
15. dawn strategy positions <conversationId>
16. dawn run logs <conversationId> --limit N
17. Stop when requested: dawn run stop <conversationId>, then verify status again.

Monitoring loop

For active monitoring sessions: 1. Query dawn run status <conversationId>. 2. Record timestamp, isRunning, status, and active strategy IDs. 3. Query dawn strategy positions <conversationId> for holdings/PnL. 4. Query dawn run logs <conversationId> --limit N for execution details. 5. If records look stale or missing, wait briefly and retry once.

Troubleshooting

• "Not authenticated. Run: dawn auth login": run dawn auth login and retry.
• Auth callback completes but CLI appears stuck: interrupt once and retry login.
• "No strategy version found...": create/revise/upload strategy code, then relaunch.
• "No strategies found for this agent" on stop: verify conversationId, then check dawn run status.
• Live launch fails: re-check funding path with dawn account fund.

Run checklist

Dawn Strategy Runbook
- [ ] Preflight complete
- [ ] Auth complete
- [ ] Funding path checked (or user confirmed paper-only)
- [ ] conversationId captured
- [ ] Strategy code generated/uploaded
- [ ] Launch run completed (paper/live)
- [ ] strategyId captured (if launched)
- [ ] Monitoring snapshots collected
- [ ] Stop executed (if requested)
- [ ] Final status verified

Skills

Individual skills for each command:

Required output

When using this skill, always return: - conversationId, - strategyId (if launched), - run mode (paper/live), - latest monitoring summary, - exact next command to run (or the last command run).