一键导入
这个仓库中的 skills
Execute one accepted task and leave verifiable evidence.
Inspect code, docs, issues, and evidence before deciding.
Lock the execution plan and task contract.
Clarify goals, constraints, boundaries, and acceptance criteria.
Continue remaining tasks, but stop when evidence is insufficient.
Execute one accepted task and leave verifiable evidence.
Inspect code, docs, issues, and evidence before deciding.
Lock the execution plan and task contract.
Clarify goals, constraints, boundaries, and acceptance criteria.
Continue remaining tasks, but stop when evidence is insufficient.
Execute one accepted task and leave verifiable evidence.
Inspect code, docs, issues, and evidence before deciding.
Lock the execution plan and task contract.
Clarify goals, constraints, boundaries, and acceptance criteria.
Continue remaining tasks, but stop when evidence is insufficient.
Execute one accepted task and leave verifiable evidence.
Inspect code, docs, issues, and evidence before deciding.
Lock the execution plan and task contract.
Clarify goals, constraints, boundaries, and acceptance criteria.
Continue remaining tasks, but stop when evidence is insufficient.
Execute one accepted task and leave verifiable evidence.
Inspect code, docs, issues, and evidence before deciding.
Lock the execution plan and task contract.
Clarify goals, constraints, boundaries, and acceptance criteria.
Continue remaining tasks, but stop when evidence is insufficient.
Execute all remaining tasks automatically after a sealed plan exists.
Inspect a repository read-only and produce a grounding handoff for seal.
Execute one sealed task through review-ready and complete.
Materialize a settled plan into plan.md and tasks.md.
Converge a rough request into a chat-only proposal-ready plan.