메뉴
Execute or implement a plan. Use only when the user explicitly invokes `execute-plan` or `$execute-plan`; do not auto-invoke from context.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
A relentless interview to sharpen a plan or design, maintain the domain language, and chart large efforts until their route is clear. Use only when the user explicitly invokes `grill` or `$grill`; do not auto-invoke from context.
Implement a piece of work based on a spec or set of tickets. Use only when the user explicitly invokes `implement` or `$implement`; do not auto-invoke from context.
Turn the current conversation into a spec and save it under .scratch/specs/ — no interview, just synthesis of what you've already discussed. Use only when the user explicitly invokes `to-spec` or `$to-spec`; do not auto-invoke from context.
Break a plan, spec, or the current conversation into tracer-bullet tickets under .scratch/, each declaring its blocking edges. Use only when the user explicitly invokes `to-tickets` or `$to-tickets`; do not auto-invoke from context.
Save a brief Architecture Decision Record (ADR) from the current session. Use only when the user explicitly invokes `adr` or `$adr`; do not auto-invoke from context.
Build extensions for Pi. Use only when the user explicitly invokes `pi-extension` or `$pi-extension`; do not auto-invoke from context.
SOC 직업 분류 기준
| name | execute-plan |
| description | Execute or implement a plan. Use only when the user explicitly invokes `execute-plan` or `$execute-plan`; do not auto-invoke from context. |
Manual invocation only: use this skill only when the user explicitly invokes execute-plan or $execute-plan; do not auto-invoke from task context.
Follow the plan to the dot and don't deviate from the plan. It's important to respect the plan during execution.
If there is an opportunity to execute tasks in parallel by spawning background agents, take that.
If a needed tool is unavailable, blocked, permission-denied, or otherwise disallowed, stop execution and explicitly ask the user for permission or direction instead of attempting an alternate workaround. This applies especially to tools that enable reading, writing, editing, deleting, or otherwise manipulating files or system state. ZERO exceptions.
These execution instructions are important. Follow them to the dot. In case you can't follow them, let the user know why.
These after execution instructions are important. Follow them to the dot. In case you can't follow them, let the user know why.
When there is an ambiguity in execution or when the execution didn't go as planned, stop the execution and present the user with options on how to proceed and ask the user for direction.
