01-initialize-srs
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill incorporates a Python script (
init_skill.py) to manage local project setup. The script performs file system operations including directory creation (Path.mkdir), copying of document templates (shutil.copy2), and conditional directory removal (shutil.rmtree) if the user selects a clean build. These actions are performed using relative paths within the project directory structure. - [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it is designed to read and process content from external markdown files supplied in the
_context/directory. - Ingestion points: Markdown files located in
projects/<ProjectName>/_context/are read to populate the SRS skeleton. - Boundary markers: The instructions include a specific 'Anti-Hallucination Guard' that explicitly forbids the agent from assuming or generating information not present in the context files, and requires tagging gaps with
[CONTEXT-GAP]. These act as effective boundary constraints. - Capability inventory: The skill has authorized capabilities to read and write files within the project repository and execute a bundled initialization script.
- Sanitization: No technical sanitization or schema enforcement is performed on the ingested markdown files beyond the instruction-based grounding requirements.
Audit Metadata