pricing-finder
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and process content from external websites (competitor product and pricing pages). This content is untrusted and could potentially contain malicious instructions intended to influence the agent's behavior during analysis. The skill uses this data to populate internal JSON files and finally generates a pricing report saved to the user's local directory.
- Ingestion points: External URLs are fetched in
SKILL.md(Step 3 and Step 6) andscripts/research.py, with content saved to/tmp/pf-product-raw.mdand/tmp/pf-pricing-raw.json. - Boundary markers: The instructions lack explicit delimiters or "ignore embedded instructions" warnings when interpolating fetched content into AI prompts.
- Capability inventory: The skill has file-writing capabilities (saving to
docs/pricing-intel/) and network access via therequestsandurllib.requestmodules inscripts/research.py. - Sanitization: There is no evidence of sanitization or filtering of the fetched external HTML/Markdown content before it is processed by the AI.
Audit Metadata