一键导入
close
Use when verified work is ready for QA, acceptance, checkpointing, learning, recommendation, and handoff.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Use when verified work is ready for QA, acceptance, checkpointing, learning, recommendation, and handoff.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Routes adaptive multi-agent deliberation with fractal context cycles. Use when using /cdo, think/deep/debug/parliament work, or long runs paired with autoresearch scheduling.
Use when you need an evidence-first provenance report from `cass`, repo docs, `.lev`, `~/.agents/diagrams`, `qmd`, or Grep/Glob tools.
Use the codebase knowledge graph for structural code queries. Triggers on: explore the codebase, understand the architecture, what functions exist, show me the structure, who calls this function, what does X call, trace the call chain, find callers of, show dependencies, impact analysis, dead code, unused functions, high fan-out, refactor candidates, code quality audit, graph query syntax, Cypher query examples, edge types, how to use search_graph.
Handle blocked destructive commands. Use when dcg blocks rm -rf, git reset --hard, DROP DATABASE, kubectl delete, or when configuring agent safety guardrails.
Use when running a visible Lev autopilot that coordinates research, interview, arch, propose, AgentPing, warm-worker intent, realtime dashboard intent, and lev exec updates across the lifecycle.
Heartbeat-driven autonomous development loop using lev loop autodev, with interval, budget, and tick controls for in-process SDLC execution.
| name | close |
| description | Use when verified work is ready for QA, acceptance, checkpointing, learning, recommendation, and handoff. |
| skill_type | workflow |
| category | lifecycle |
| output_template | hud |
| triggers | ["close","ok","done","accept"] |
Close the active entity: capture leftovers, verify acceptance, checkpoint, learn, recommend, and hand off.
Lifecycle lane: Close
Entity movement: verified -> closed | monitoring | follow_up | blocked
Workstream: update active workstream before reporting done
Upstream: /exec
Downstream: /handoff, /work
Router: /work
HUD: end with 🧬 {ws} ⚡{exec_count} 📥{capture_count} ⏸️{paused_count} ✅{done_count} | 🚦{gate}={score} | ⏭️ {next} | 🔁{loop_state}
steps:
- id: capture_gate
action: Run capture in gate mode before closing.
validation: "in_memory_count == 0."
on_failure: "Route to /capture and stop close."
- id: qa_acceptance
action: Check output against task DNA, acceptance criteria, declared verifier evidence, and proof-gate verdicts.
validation: "Every acceptance item and applicable proof gate is pass, fail, blocked, or explicitly not-tested with residual risk."
on_failure: "Route failed criteria back to /exec."
- id: proof_gate_verdict
action: Seal Pentagon, UltraQA, and ai-slop-cleaner outcomes before acceptance.
validation: "Feature-local Pentagon verdicts, UltraQA scenario matrix status, generated-artifact cleanup, ai-slop-cleaner review, and residual risks are recorded."
on_failure: "Do not accept; route to /exec or /propose depending on whether runtime evidence or proof contract is missing."
- id: accept
action: Decide whether the work is accepted.
validation: "Confidence >= 0.85 for auto-accept, otherwise human approval is recorded."
on_failure: "Do not commit or handoff as closed."
- id: checkpoint
action: Stage, commit, pull without rebase, and push when sealing verified work.
validation: "Commit and push succeed, or blocker is recorded."
on_failure: "Keep entity blocked and do not claim closed."
- id: learn
action: Extract typed learnings with claim, evidence, applies_to, and propagates.
validation: "At least one learning or explicit 'no new learning' is recorded."
on_failure: "Add the learning block before handoff."
- id: recommend
action: Recommend keep, monitor, execute, or reject for related entities.
validation: "Next lifecycle verb is clear."
on_failure: "Route to /work for next-verb selection."
- id: handoff
action: Invoke /handoff to update workstream state and resume prompt.
validation: "Workstream YAML reflects close result."
on_failure: "Do not end without durable state."
When the workstream has a trigger stream, write:
schema: lifecycle_trigger.v1
verb: close
entity_ref: ws://<ws-id> | task://<task-id>
outputs:
items: []
learnings: []
gates: []
proof_gates:
pentagon: pass | fail | blocked | not_applicable
ultraqa: pass | fail | blocked | not_applicable
ai_slop_cleaner: pass | fail | blocked | not_applicable
artifacts: []
metrics:
drift: <number>
in_memory_count: 0
cursor:
last_verb: close
next_verb: null
recommend: []
| KPI | Move on |
|---|---|
| drift_score | < 0.3 |
| hygiene_pass_rate | >= 0.95 |
| learning_extracted | >= 1 or explicit none |
| proof_gate_verdicts | all applicable gates pass or have explicit blocked/follow-up routing |
Max rework: 2. Three consecutive gate failures becomes blocked.
/exec provides verified work./capture clears in-memory leftovers./handoff updates workstream continuity./work routes the next entity.