ops-orchestrate
Autonomous multi-project orchestration engine. Audits all registered projects, structures work into dependency-wired tasks, dispatches parallel agents (subagents or Agent Teams), audits completions, and ships PRs. Registry-driven — works for any user with a configured project registry.
npx skills add davepoon/buildwithclaude --skill ops-orchestrate --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Audits all registered projects, structures work into dependency-wired tasks, dispatches parallel agents (subagents or Agent Teams), audits completions, and ships PRs. Registry-driven — works for any user with a configured project registry.
How it works
- Loads runtime context (preferences, daemon health, secrets, memories) before orchestration.
- Phase 1 — Audit: discovers projects (SCOPED or GLOBAL), dispatches parallel read-only audits in batches, collecting git status, unpushed commits, open PRs, recent CI runs, planning state, roadmaps, and unresolved PR reviews; runs external issue queries in parallel; cross-references with existing tasks.
- Phase 2 — Structure: decomposes work into atomic tasks (one PR max, one repo per task) with clear acceptance criteria and rich metadata; wires dependencies to maximize parallelism (no unnecessary serialization); assigns waves (0, 1, ...); tags tasks for subagent or team mode in hybrid.
- Phase 3 — Dispatch: max parallelism dispatches up to 5 agents (subagents or teams), each with isolated worktrees; uses model: "sonnet"; run_in_background: true; marks tasks in_progress before dispatch; applies appropriate Agent/Team prompts.
- Phase 4 — Audit: upon completion, verifies PR diffs against task, runs quality gates, checks CI, conducts security reviews if touched, and re-dispatches or messages as needed.
- Phase 5 — Ship: resolves reviews, waits for CI, handles merges (squash/admin), and performs post-deploy verifications.
- Phase 5.5 — Liveness Monitor: monitors in-flight agents, handles stalled agents with targeted re-dispatch or stop, avoiding re-dispatching identical prompts.
- Phase 6 — Loop: continuously cycles tasks until completion or blockers surfaced; prints final report when done.
## Runtime Context Before orchestrating, load: 1. **Preferences**: `cat ${CLAUDE_PLUGIN_DATA_DIR:-$HOME/.claude/plugins/data/ops-ops-marketplace}/preferences.json` — read `owner`, `timezone`, `yolo_enabled`, registry path 2. **Daemon health**: `cat ${CLAUDE_PLUGIN_DATA_DIR}/daemon-health.json` — ensure all services healthy before dispatching 3. **Secrets**: Resolve via env → Doppler → password man
What does the ops-orchestrate skill do?
Autonomous multi-project orchestration engine. Audits all registered projects, structures work into dependency-wired tasks, dispatches parallel agents (subagents or Agent Teams), audits completions, and ships PRs. Registry-driven — works for any user with a configured project registry.
How do I install it?
Run `npx skills add davepoon/buildwithclaude --skill ops-orchestrate --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From davepoon/buildwithclaude, a repository with 3,251 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.
