do-execute-all-tasks
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes metadata (task IDs and directory names) from an external
tasks.mdfile and interpolates this data into prompts sent to a high-capability subagent (agent-execute-task). - Ingestion points: The orchestration logic reads
tasks.mdusing file system tools likeread,glob, orlist_dirdepending on the platform. - Boundary markers: The skill does not use explicit boundary markers or XML-style delimiters to isolate interpolated variables (like
<ID>or<nome>) from the rest of the subagent prompt. - Capability inventory: The targeted subagent (
agent-execute-task) is granted broad permissions, including reading specifications, modifying code, and executing tests. - Sanitization: There is no evidence of validation or sanitization of the strings extracted from the task file before they are used to build the delegation command.
Audit Metadata