plan
Pass
Audited by Gen Agent Trust Hub on Apr 10, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements robust planning protocols and quality gates, emphasizing code understanding and iterative verification.
- [COMMAND_EXECUTION]: The skill uses platform-native tools like
TaskCreateand triggers internal slash commands (e.g.,/plan-hard,/plan-fast). These operations are restricted to the agent's internal orchestration framework and do not involve arbitrary shell execution. - [PROMPT_INJECTION]: The skill processes user-provided task descriptions through the
$ARGUMENTSplaceholder. While this presents an attack surface for indirect prompt injection, the risk is mitigated by the use of<task>boundary markers and the requirement for user confirmation and validation (AskUserQuestion,/plan-validate,/plan-review) at critical decision points. - [DATA_EXPOSURE]: The skill reads project documentation and source code to generate informed plans. This activity is consistent with the skill's stated purpose and does not involve accessing sensitive system credentials or exfiltrating data to external domains.
Audit Metadata