orchestrating-subagents — Claude Code skill: setup and usage
orchestrating-subagents is a Claude Code skill for dividing project tasks among subagents, with parallel explorers and independent code review in isolated contexts.
MIT ★ 539 updated 2026-10-29
npx skills add ZaxbyHub/opencode-swarm --skill orchestrating-subagents -a claude-codeWhat orchestrating-subagents does
This skill establishes economics and discipline for delegating tasks across subagent roles. Explorers can run on lighter models with lower effort settings because downstream review catches omissions. In contrast, reviewers and critics must use the strongest session model with high effort. They always launch in fresh contexts so prior conversation history never biases their code validation.
Parallel fan-out requires disjoint work boundaries. The orchestrator must fire independent agents within a single prompt, avoiding overlapping directories or redundant file inspections. Every delegation prompt specifies the deliverable structure, concrete file:line citation requirements, and rigid outcome labels. Heavy search outputs stay encapsulated inside subagents, which keeps the primary context uncluttered.
Direct execution handles single-fact lookups that need only one or two Grep or Read calls. Spawning overhead makes sense only when buying breadth or independent verification.
What it helps with
- Assign subagent roles with dedicated model tiers for discovery, implementation, and critique
- Run independent code reviews using fresh agents with no prior conversational bias
- Restrict subagent prompt scopes and require precise file and line evidence in results
- Keep the orchestrator context uncluttered by offloading heavy reading tasks
Example prompts for the agent
- Launch parallel subagents to map out payment webhook handlers in
src/billing/without overlapping scopes - Spin up an adversarial reviewer in a fresh context to evaluate the latest diff against our conventions
- Draft a bounded prompt for an explorer agent to locate unused database models and report file:line references
How to install orchestrating-subagents
1Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill orchestrating-subagents -a claude-code2Manually: put the files in place
~/.claude/skills/orchestrating-subagents/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/orchestrating-subagents/
1Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill orchestrating-subagents -a codex2Manually: put the files in place
.agents/skills/orchestrating-subagents/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill orchestrating-subagents -a cursor2Manually: put the files in place
.cursor/skills/orchestrating-subagents/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install ZaxbyHub/opencode-swarm .claude/skills/orchestrating-subagents2Manually: put the files in place
.github/skills/orchestrating-subagents/SKILL.md1Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill orchestrating-subagents -a gemini-cli2Platform CLI
gemini skills install https://github.com/ZaxbyHub/opencode-swarm --path .claude/skills/orchestrating-subagentsManually: put the files in place
.devin/skills/orchestrating-subagents/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add ZaxbyHub/opencode-swarm --skill orchestrating-subagents -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about orchestrating-subagents
How to install orchestrating-subagents in Claude Code?
Run npx skills add ZaxbyHub/opencode-swarm --skill orchestrating-subagents -a claude-code in your terminal.
Which subagent roles can use smaller models?
Down-tiering model selection and reasoning effort is allowed only for explorer roles that gather candidates. Reviewers and critics require the primary session model running at high effort in a fresh context.
What happens if the harness cannot spawn subagents?
The agent runs the review checklist directly as an explicit fallback self-review and notes in the final report that an isolated reviewer was unavailable.
Source text
SKILL.md · MIT
--- name: orchestrating-subagents description: > Tiering and economics for delegating to subagents: which agent type, model, and effort to use per role (explorer, implementer, reviewer, critic), how many agents to launch in parallel, how to write scoped subagent prompts with bounded structured returns, and how to keep the main context clean. Use when launching subagents, parallel explorers, independent reviewers, or critic passes — especially for swarm-mode, qa-sweep, or issue-tracer work. metadata: audience: swarm-plugin --- # Orchestrating Subagents Swarm-mode work in this repo delegates heavily (explorer → reviewer → critic). This skill defines HOW to delegate so validation gates stay strong while breadth stays fast and cheap. It complements — never replaces — the gates in `.claude/session/swarm-mode.md`, qa-sweep, and swarm-implement. ## Role → tier mapping | Role | Agent type | Model / effort | Rationale | |---|---|---|---| | Explorer (mapping, candidate findings) | `Explore` when read-only suffices | Cheaper/faster tier acceptable; low–medium effort | Recall-bound, not reasoning-bound; the reviewer gate catches misses | | Implementer (scoped edits) | general-purpose | Session model; medium–high effort | Edits need project conventions (CLAUDE.md context) | | Reviewer (independent validation) | general-purpose, **fresh context** | Session (strongest) model; high effort | Precision-bound; false approvals are the expensive failure | | Critic (final challenge) | general-purpose, **fresh context** | Session (strongest) model; high effort | Same — this is the last line of defense | Hard rule: economize on explorers, never on reviewers or critics. If the harness exposes model or effort overrides for subagents, tier explorers down; do not tier the reviewer or critic below the session model or below high effort. If no override is available, tier by agent type (`Explore` is lightweight: read-only tools, skips CLAUDE.md) and by prompt scope. ## Fan-out discipline - Launch parallel agents only for **disjoint scopes**. Before launching, write one line per agent stating its scope; if two overlap, merge them. - 2–4 explorers per wave is a heuristic only when the active workflow does not define its own fan-out contract. A workflow-specific contract overrides this heuristic: PR review, for example, requires all six base dimensions and all eleven risk families to be covered on every PR — its capability profiles and depth tiers (see swarm-pr-review) govern how many lanes carry that coverage, and under the plugin's mechanical controller the tier is computed from the bound diff and the matching lane floors are enforced outright. No time, cost, repository-size, or simplicity rationale may reduce the coverage a workflow mandates. Outside fixed-fan-out workflows, more agents than distinct scopes adds token cost and synthesis burden without adding recall. - Launch independent agents **in a single message** so they run concurrently. - Do not re-run a search an agent is already doing; wait for its report. - Scale waves, not width: if the first wave surfaces new territory, launch a second targeted wave rather than one giant speculative first wave. ## Subagent prompt contract Every delegation prompt must state: 1. **Scope** — exact directories, files, or question. Name what is out of scope. 2. **Deliverable** — the structure of the report (per-item findings, then a ranked summary). The agent's final message is the only thing you receive. 3. **Evidence bar** — exact `file:line` references; no invented paths; verify a path exists before citing it. 4. **Status labels** — explorers return CANDIDATE findings only; reviewers classify CONFIRMED / DISPROVED / UNVERIFIED / PRE_EXISTING; reviewer and critic verdicts are APPROVE / NEEDS_REVISION / BLOCKED. 5. **Output bound** — compact structured returns; no full-file dumps. For reviewers and critics additionally: - Give the **claims and locations**, not the author's justification — the reviewer must re-derive, not confirm, the reasoning. - State the adversarial default explicitly: "default to DISPROVED/UNVERIFIED unless the code evidence supports the finding." - A reviewer or critic must be a **fresh agent**, never a continued conversation with the agent whose work it judges. ## Independence and staleness - Reviewer and critic review the **latest diff**, not a description of it. Give them the branch state and the validation evidence, and require them to run `git diff`/`git log` themselves. - Any edit after an approval invalidates it. Record what was approved (e.g. `git rev-parse HEAD`, `git diff --stat`) so staleness is checkable — see the durable-session-state skill. ## Nesting limitation Whether a subagent can spawn further subagents depends on the harness and agent type — check whether a subagent tool (`Agent` or `Task`) is actually available in your context before assuming either way. If a skill mandating fresh-subagent review (qa-sweep, swarm-implement) executes in a context **without** a subagent tool: - perform the same review checklist yourself as a clearly labeled **fallback self-review**, and - disclose in your report that independent review was unavailable in this context, so the orchestrator can re-run the gate with a real fresh agent. Never silently present self-review as independent review. ## Main-context hygiene - Push reading-heavy work into subagents; keep the main thread for scoping, synthesis, and decisions. - Do not paste subagent transcripts or large tool outputs back into the main thread; carry forward only validated findings and verdicts. - When a subagent report arrives, extract the load-bearing facts into your durable task artifacts (see durable-session-state) before moving on. ## When not to delegate Answer directly, without a subagent, when the task is a single-fact lookup you can resolve with one or two targeted Grep/Read calls, or when you already know the file and symbol. Delegation overhead should buy breadth, isolation, or independence — if it buys none of those, skip it.
© ZaxbyHub · MIT · ZaxbyHub/opencode-swarm