exec-plans
Pass
Audited by Gen Agent Trust Hub on Mar 27, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains instructions that explicitly mandate the overriding of the agent's default operational constraints, particularly regarding conciseness and document structure.\n
- Evidence: 'When generating the final plan, the rules here OVERRIDE Plan Mode's default constraints'\n
- Evidence: 'Ignore Plan Mode's instructions to be "concise by default"'\n- [COMMAND_EXECUTION]: The skill encourages the agent to operate with high autonomy and execute shell commands (running the project and tests) while suppressing user intervention between steps.\n
- Evidence: 'do not prompt the user for "next steps"; simply proceed to the next milestone'\n
- Evidence: 'Resolve ambiguities autonomously'\n
- Capability: The skill explicitly grants the agent permission to 'run the project, and run tests' based on the generated plan without human-in-the-loop validation of the specific commands.
Audit Metadata