메뉴
Use when the real risk is not lack of knowledge but self-justifying shortcuts such as explaining instead of running commands, skipping the hard part first, or deciding on your own that verification would take too long
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
Use when verifying implementation work and the failure mode is superficial approval, code-reading in place of execution, or over-trusting a passing test suite without trying to break the change
Use when coordinating multi-step work across workers or subagents, especially when research, implementation, and verification should be split while the main agent stays responsible for synthesis and user communication
Use when the task is codebase exploration rather than modification, and you need a fast, read-only search pattern that starts broad when location is unknown, narrows when location is known, and parallelizes independent searches
Use when deciding what information belongs in memory, how to classify it, and what must not be saved because it is derivable from the repo, git history, or current task context
Use when deciding whether shell commands are truly read-only, safe for concurrent execution, and free of path, git, redirection, or command-shape risks that should force a stricter permission path
Use when maintaining session notes that must preserve a fixed markdown template, stay within section and total token budgets, and remain useful after compaction or resume
SOC 직업 분류 기준
| name | self-rationalization-guard |
| description | Use when the real risk is not lack of knowledge but self-justifying shortcuts such as explaining instead of running commands, skipping the hard part first, or deciding on your own that verification would take too long |
This skill is a stop-loss against fake progress. It is for moments when you can feel yourself drifting from execution into explanation.
It is informed by the anti-rationalization rules in the built-in verification agent and the discipline-oriented superpowers skills.
When you hear yourself making an excuse, translate it into the next concrete action.
Do not debate the excuse. Replace it.
| Excuse | Required action |
|---|---|
| "The code looks correct." | Run it. |
| "The tests probably cover it." | Verify the specific behavior yourself. |
| "This would take too long." | State the estimated time and do it anyway. |
| "I'll do the easy part first." | Start with the hardest or riskiest part. |
| "Let me explain what I would check." | Stop writing and run the command. |
| "This is probably unrelated." | Gather evidence before dismissing it. |
Stop immediately if you are:
If a task has one hard edge and several easy edges, do the hard edge first.
Examples:
Easy progress is often camouflage for avoiding the real risk.
Prefer observable actions over speculative prose.
Good:
Bad:
If the correct action is expensive, say so plainly, then do it.
Do not silently downscope the task because you think it is too slow, too annoying, or too involved.
You have applied this skill correctly when each moment of self-justification is converted into an executed check, a concrete command, or a direct attempt on the hardest remaining risk.
