specstory-link-trail
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes session history markdown files which contain content from previous web fetches and tool interactions. This content could potentially include instructions designed to influence the agent when it analyzes the generated report. \n
- Ingestion points: The script
parse_webfetch.pyreads history files from the.specstory/history/directory. \n - Capability inventory: The skill uses
BashandReadto execute Python scripts that parse local files and output markdown. \n - Boundary markers: The agent is not provided with explicit boundary markers or instructions to ignore embedded instructions in the generated report. \n
- Sanitization: The report generation logic extracts raw text segments (such as titles and summaries) from the history files with minimal sanitization, potentially preserving malicious payloads.
Audit Metadata