break-down-prd
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes an internal Python helper script,
scripts/skillsrc.py, usinguv run. This execution is limited to managing the skill's specific configuration within the.draekien/.skillsrcJSON file and does not interact with external systems or broader environment settings.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes product requirements documents (PRDs) provided by the user, which represents a potential surface for indirect instruction injection.\n - Ingestion points: PRD content is read from a user-specified file path or the current session context within
SKILL.md.\n - Boundary markers: The skill does not implement specific delimiters or instructions to ignore embedded commands within the PRD input.\n
- Capability inventory: The skill can write markdown files to the local filesystem and execute a specific internal configuration script.\n
- Sanitization: The skill maps input content into structured markdown templates but does not perform explicit sanitization or filtering of the text.\nSince the skill is intended for planning and produces static documentation as its primary output, the risk of these inputs causing malicious behavior is minimal.
Audit Metadata