run
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill provides logic through natural language instructions and YAML configuration. No executable scripts, binaries, or third-party packages are included.
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because its control flow is driven by the contents of
project.md. Maliciously crafted content in that file could potentially influence the selection of tasks or inputs passed to child skills. - Ingestion points: Reads project state from
project.mdin Step 1. - Boundary markers: No delimiters or explicit instructions to ignore embedded commands are present in the procedure.
- Capability inventory: Dispatches execution to other skills including
asta-assistant:do-work,asta-assistant:plan-work,asta-assistant:brainstorm, andasta-assistant:save-work. - Sanitization: No explicit validation or sanitization of the parsed project data (slugs or statuses) is performed before use.
Audit Metadata