market-research-reports
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources using the research-lookup tool to gather market statistics and trends.
- Ingestion points: Research data is gathered in SKILL.md (Phase 1, Step 2).
- Boundary markers: The instructions lack explicit boundary markers or warnings to the agent to ignore instructions embedded within the fetched research data.
- Capability inventory: The skill has Bash access, performs file write operations, compiles LaTeX documents, and executes sub-processes via Python.
- Sanitization: There is no evidence of sanitization or filtering of the external content before it is processed or included in report drafts.
- [COMMAND_EXECUTION]: The skill programmatically executes local Python scripts to generate report visualizations.
- Evidence: The file scripts/generate_market_visuals.py uses subprocess.run to call generate_schematic.py and generate_image.py.
- Details: User-supplied input (the --topic argument) is interpolated into the prompts passed as arguments to these sub-processes. While subprocess.run is used with a list (reducing shell injection risk), the resulting data flow relies on the downstream scripts' safety.
Audit Metadata