본문으로 건너뛰기
cccskills
무료GitHub에 공개

dispatching-parallel-agents

The parallel fan-out primitive. Routed to by any skill or command that splits work across independent units and dispatches an agent per unit — subagent-driven-development, /sprint, parallel /review. It owns the dispatch/collect/funnel discipline: bound concurrency, isolate units, collect every result, dedupe overlap, and funnel through finding-triage then the read-only verdict-aggregator. Raw agent output is never consumed before the funnel runs; an agent that errors records its unit as incomplete without corrupting the batch.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

설치 방법 보기

포함된 파일(1)

  • SKILL.md6.2 KB

SKILL.md(원문)

설치하기 전에 에이전트에게 전달되는 지시 내용을 확인할 수 있습니다.

dispatching-parallel-agents

Fan out, collect everything, funnel before consuming. Plumbing other skills dispatch through — it owns the parallel discipline, not domain logic.

Pre-flight

The caller supplies the work. Confirm it, or STOP and surface the gap — never invent units or pick an agent on a hunch:

  • The unit list — N discrete, independently-completable units of work, each with its scope and target paths.
  • The agent kind — which agent kind to dispatch per unit (e.g. backend-author, frontend-author, scout, security-reviewer at {{PLUGIN_ROOT}}/agents/<name>.md). One kind per batch.
  • The completion contract — what each agent must return: a structured result, its unit ID, and a pass/fail/error status. Self-reports are claims, not evidence (see Phase 4).

Phase 1 — Partition · gate: BLOCK

Split the work into units that do not collide. A unit owns a distinct file or path set; two units in one batch MUST NOT mutate the same path.

  • Independent paths — units touch disjoint files. Dispatch directly. This is the common, fast case.
  • Shared paths unavoidable — units must mutate the same file or tree. Do NOT dispatch into the conflict. Route to using-git-worktrees ({{PLUGIN_ROOT}}/skills/using-git-worktrees/SKILL.md) to give each unit an isolated working tree, or serialize the colliding units into one sequential unit. A shared-path collision in a parallel batch is a corruption, not a merge.

Tag each unit with its ID, scope, target paths, and the agent kind. Read-only batches (review, scout) skip the collision check — they mutate nothing.

Gate: every unit has a disjoint path set, or its overlap is resolved by worktree isolation or serialization. An unresolved collision does not pass.

Phase 2 — Dispatch · gate: BLOCK

Dispatch one agent per unit, bounded.

  • Concurrency bound — dispatch in waves of at most a fixed width (default 4 concurrent agents; the caller MAY override). Do not fan out unbounded; a hundred agents at once starves the batch and corrupts nothing useful.
  • Fresh and isolated — each agent gets only its own unit's scope and the completion contract. No agent sees another's working state. No shared mutable context between units in flight.
  • Hard-stop propagation — a tdd BLOCK, a commit-gate failure, a security CRITICAL, or a [CONFIRM-NN] inside any unit halts that unit and surfaces immediately; it does not get silently swallowed by the batch.

Gate: every unit is dispatched or explicitly deferred. No unit is dropped silently at dispatch.

Phase 3 — Collect · gate: BLOCK

Wait for the full wave, then the next, until every unit has terminated. Record each unit's terminal state: DONE (returned a result), ERRORED (the agent failed), or DEFERRED (a hard-stop surfaced).

An ERRORED unit drops its own work — its partial output is discarded, never half-applied — and is logged against its unit ID. One errored unit MUST NOT corrupt or discard the batch. The remaining units' results stand. Record the error for the funnel; do not retry blindly inside this phase.

Gate: every unit has a terminal state and DONE results are captured intact. A still-running or unaccounted unit blocks Phase 4 — partial collection does not pass.

Phase 4 — Dedupe · gate: BLOCK

Parallel agents over adjacent units produce overlapping findings — the same issue reported by two units, the same fix proposed twice. Collapse them before anything downstream sees the batch.

  • Merge findings that name the same defect at the same location into one, keeping the highest severity reported.
  • Reconcile contradictory results across units (two units that disagree on the same fact) into a single surfaced conflict, not a silent pick.
  • Verify any "done" claim that gates real work: re-run the proving command fresh and read its output and exit code. A subagent's self-report is a claim — never trust it as evidence of completion.

Gate: the result set is deduped, contradictions surfaced, and completion claims verified by fresh evidence. Raw, un-deduped agent output MUST NOT reach Phase 5.

Phase 5 — Funnel · gate: STOP

The batch is consumed only here, through the fixed funnel — never directly by the caller.

  1. Dispatch finding-triage ({{PLUGIN_ROOT}}/agents/finding-triage.md) over the deduped result set and batch completion contract: it classifies severity and marks out-of-scope items with an inline [NEEDS-TRIAGE] marker. Every reviewer finding reaches the verdict; only non-finding transport metadata may be omitted.
  2. Hand the triaged set to verdict-aggregator ({{PLUGIN_ROOT}}/agents/verdict-aggregator.md): it composes the single read-only batch verdict the caller consumes — PASS, BLOCKING_FINDINGS, or INCOMPLETE.

The errored and deferred units from Phase 3 ride through the funnel as findings — an ERRORED unit is a finding the caller must see, not a silent gap.

Gate: the caller receives only the aggregated verdict. Bypassing the funnel — consuming raw or merely-deduped agent output — is a gate violation. STOP and return the verdict.

Hard rules

  • MUST NOT dispatch two units that mutate the same path in one batch; isolate via using-git-worktrees or serialize.
  • MUST NOT fan out unbounded; dispatch within the concurrency bound.
  • MUST NOT let one ERRORED unit discard or corrupt the rest of the batch.
  • MUST NOT consume agent output before the finding-triage → verdict-aggregator funnel runs.
  • MUST NOT trust a subagent's self-reported completion; verify with a fresh proving command.
  • MUST NOT silently drop a unit — every unit terminates with a recorded state that rides through the funnel.

리뷰

아직 리뷰가 없습니다. 사용해 본 소감을 남겨 주세요.

댓글

로그인 없이 쓸 수 있습니다. 사용 후기, 질문, 작동하지 않은 환경 등을 편하게 남겨 주세요.

아직 댓글이 없습니다. 첫 댓글을 남겨 보세요.

보내면 바로 공개됩니다. 이메일 주소나 전화번호 같은 개인 정보는 쓰지 마세요. URL은 2개까지입니다.

0 / 1,000자

이 양식은 reCAPTCHA로 보호되며 Google 개인정보처리방침 및 서비스 약관이 적용됩니다.

같은 저장소의 스킬

Refine feature and sprint intent into a useful specification. Capture scope, observable criteria and material choices without repeated artifact approval ceremonies.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

arbiterForge/codeArbiter1492026년 10월 20일 업데이트

Vet a new or changed third-party dependency for license, provenance, and supply-chain risk before any install runs.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

arbiterForge/codeArbiter1492026년 10월 20일 업데이트

ca-adr

무료

Record user-decided ADRs or inspect their health read-only. Preserve attribution and acceptance evidence.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

arbiterForge/codeArbiter1492026년 10월 20일 업데이트

Inspect ADR health read-only; optionally select one ADR with --adr N.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

arbiterForge/codeArbiter1492026년 10월 20일 업데이트

ca-audit

무료

Assemble the governance record for a range — commits, overrides, ADRs, sprint auto-decisions, open questions, checkpoint findings — into one dated audit packet. Read-only.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

arbiterForge/codeArbiter1492026년 10월 20일 업데이트

ca-btw

무료

Lightweight Q&A about the project — answer from context and return, no routing, no state change.

한국어 요약은 준비 중입니다. 원문 설명을 표시하고 있습니다.

arbiterForge/codeArbiter1492026년 10월 20일 업데이트

arbiterForge의 스킬 모두 보기

이 스킬의 문제 신고