La interfaz está localizada; las descripciones y documentos de las Skills se traducen automáticamente con IA a su idioma. El código, los comandos, rutas y términos técnicos se mantienen en el original, y prevalece el documento del autor.
Explorar Skills
Productividad · SKILL.md

subagents

subagents

Use when a plan has parallel tasks and the harness supports dispatching subagents.

Independent GitHub snapshot · Automated classification · Original documentation governs

Documentación de la Skill

GitHub

name: subagents description: Use when a plan has parallel tasks and the harness supports dispatching subagents.

Subagents

If topology=multi-module (triage announcement or coordinator spec/plan declaration), read ../references/multi-module.md.

Fresh context per task. No session history.

Dispatch: reference, don't copy

Subagents have file-read tools. Don't inline what they can read.

TASK: <one sentence>
ROLE: <implementer | spec-reviewer | quality-reviewer>
MODEL: <fast | balanced | strongest>
REPOSITORY: <registered repo id and path>
MODULE: <module id and root>
FILES: <paths>
SPEC: <docs/staging/specs/...#anchor>
ACCEPTANCE: <test or cmd>
EXTRA: <only what's NOT in referenced files>

ROLE is required. See role charters below for what each role owns and does not own.

MODEL is optional. Omitted → uses harness default. Resolution: .praxis/model-tiers.yaml in project root or user home maps the three tiers (fast, balanced, strongest) to concrete model IDs. No config file → all subagents use harness default.

Pasting >10 lines from the spec? Stop - let the subagent read it.

Expand milestone → atomic steps at dispatch time, not in the plan.

One task has one repository and module write scope. Agents may read shared coordinator contracts, but must not edit another module or repository. Never dispatch two writers to overlapping module roots. Integration tasks are dispatched only after every required module reports local acceptance.

Loop per task

implementer → on DONE: spec-reviewer (matches spec?) → quality-reviewer (review) → mark complete, continue. The coordinator marks - [x], never the subagent.

Status

  • DONE_WITH_CONCERNS - address if correctness; note if observation.
  • NEEDS_CONTEXT - supply the fact, re-dispatch.
  • BLOCKED - diagnose (missing context / stronger model / too big / plan wrong). Never silently retry.

Model tier

MODEL: fast | balanced | strongest — optional, omitted → harness default.

Resolved via .praxis/model-tiers.yaml in project root or user home. Template at Praxis repo root: model-tiers.example.yaml.

The coordinator decides the tier at dispatch time based on task complexity — it is not hardcoded per role. The role charters below list defaults; override when:

ScenarioTierExamples
Mechanical edits, simple checksfastrename a constant, fix a typo, pattern-match review
Standard implementation, single-file featurebalancedadd a function, modify one module, write routine tests
Multi-file refactor, concurrency/state-machine, safety reviewstrongestimplement a protocol, review crash-recovery, cross-module integration

Role charters

implementer

  • Owns: writing the minimum code to make the acceptance test pass (GREEN phase). This includes writing the failing test first (RED), then the implementation (GREEN), then refactoring within the changed files.
  • Does NOT own: expanding scope beyond the task, adding abstractions not required by the current test, touching files outside FILES, marking the task complete.
  • Default tier: balanced. Override to fast for mechanical edits, strongest for multi-file protocol/state-machine implementation.
  • On DONE: hands off to spec-reviewer.

spec-reviewer

  • Owns: verifying implementation matches every contract and acceptance criterion in SPEC. Reports drift with file:line references.
  • Does NOT own: suggesting improvements beyond the spec, checking code quality, re-running tests.
  • Default tier: fast. Override to balanced when spec has ≥3 cross-referencing rules or state enums.
  • On DONE_WITH_CONCERNS: coordinator decides whether to address or note.

quality-reviewer

  • Owns: running review skill against the diff — convention compliance, simplification opportunities, idiomatic patterns, security, scope.
  • Does NOT own: re-verifying spec compliance (already done), re-running acceptance tests (already done).
  • Default tier: strongest. Override to balanced for single-file changes where deep reasoning is unnecessary.
  • On DONE: task is complete. Coordinator marks - [x].

All done → review whole diff → ship.

Skills guardadas