Menú
Decompose a BRD into epics, user stories, acceptance criteria, and a dependency graph with parallel groups for agent team execution.
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Basado en la clasificación ocupacional SOC
Switch model mid-session preserving thinking blocks. Used by /model command and by BRD §6.2 failover when the primary provider rate-limits or fails.
Six-phase ReAct loop (pre-check, thinking, self-critique, action, tool, post). Activates per-workflow via the thinking_level knob in config/workflows.yaml. Replaces standard ReAct for workflows where independent verification of the reasoning improves outcomes.
Mines completed sessions for repeating {tool sequence → outcome} tuples. Scores by frequency × success_rate × novelty. High-scoring tuples become "instincts" — candidate skill seeds — in instincts/pending/.
Walk-back algorithm for the Spec-Auditor subagent. Given a failure at phase N, identifies the earliest phase whose spec, if tightened, would have prevented the failure. Proposes a surgical amendment.
Sessions stored as trees (not lists). /fork creates a branch from any point; /tree navigates; /branch labels a path; /export produces HTML for review. All branches live in one session file under sessions/<project>/<session_id>.json.
Progressive context refinement for subagents. Don't dump everything; retrieve in passes. Start with a list of file paths and one-line abstracts; load full content only when the abstract proves insufficient.
| name | spec |
| description | Decompose a BRD into epics, user stories, acceptance criteria, and a dependency graph with parallel groups for agent team execution. |
| disable-model-invocation | true |
| argument-hint | [path-to-BRD] |
/spec [path-to-BRD]
Read .claude/skills/spec-patterns/SKILL.md for decomposition patterns.
Spawn spec-writer agent with the BRD as input.
specs/brd/changelog.md or specs/brd/app_spec.md frontmatter) to the agent. Stories must include BRD Version: v{N} in their frontmatter.Agent writes output to specs/stories/:
epics.md — epic summaries with story listsE{n}-S{n}.md — individual story files with acceptance criteriadependency-graph.md — hard dependencies + parallel groupsGenerate features.json in the project root. CRITICAL FORMAT: Must be a JSON array (not a dict/object). One entry per story. The /auto loop reads this file with json.load() and iterates it as a list. Wrong format = auto loop breaks.
Story file naming: Use E{n}-S{n}.md format (e.g., E1-S1.md, E3-S4.md). Alternative: {GROUP}-{NN}-{slug}.md (e.g., A-01-types.md). The id field in features.json must match the filename stem.
Schema per entry (all 10 fields required):
[
{
"id": "E1-S1",
"category": "foundation",
"story": "Define shared types and enums",
"group": "A",
"description": "One-line summary",
"steps": ["AC1: Given...", "AC2: Given..."],
"passes": null,
"last_evaluated": null,
"failure_reason": null,
"failure_layer": null
}
]
Validation: After writing, verify: python3 -c "import json; d=json.load(open('features.json')); assert isinstance(d, list), 'Must be array'; print(f'{len(d)} features')"
Verify gate: every story has acceptance criteria, layer assignment, and group.
Present epic summary and dependency graph for user review.