paw-ps-research
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from web searches and page crawls, creating an attack surface for indirect prompt injection.\n
- Ingestion points: Data is ingested via
web_search_exaandcrawling_exatools used for competitor analysis, demand signals, and gap identification as defined in the 'Public Web Research' section of SKILL.md.\n - Boundary markers: The instructions emphasize 'evidence-first' reasoning and being 'skeptical by default,' but no explicit delimiters for raw web content are specified.\n
- Capability inventory: The agent can write synthesized research and logs to local files in the
.pawbytesdirectory structure, as described in the 'What to Write' section of SKILL.md.\n - Sanitization: There is no mention of sanitizing or escaping content retrieved from external web sources before it is processed by the agent.
Audit Metadata