one-lesson-rule
Pass
Audited by Gen Agent Trust Hub on Nov 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONTIME_DELAYED_CONDITIONALPERSISTENCE
Full Analysis
- [COMMAND_EXECUTION]: The
scripts/sentinel.mjstool usesexecSyncandspawnSyncto run multiple shell commands. These primarily consist ofgitoperations likelog,status, anddiffused for gathering project statistics. - [DYNAMIC_EXECUTION]: The skill dynamically resolves and executes a secondary script,
check-plan.mjs, from a path relative to the skill's own script directory. - [INDIRECT_PROMPT_INJECTION]: The sentinel script reads data from the project repository, such as commit messages and code changes. This content is then summarized and presented to the AI agent, creating a surface where malicious instructions in the codebase could influence agent behavior.
- [TIME_DELAYED_CONDITIONAL]: The sentinel mechanism is designed as a background loop that activates at specific time intervals derived from the project's time budget.
- [PERSISTENCE]: The skill guides the agent to use platform features like
CronCreateor background shell loops to schedule recurring execution of the monitoring script, ensuring the monitoring process persists throughout the development phase.
Audit Metadata