side-hustle-maker
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the web through specialized sub-agents. * Ingestion points: Market research data fetched using WebSearch and WebFetch from sites like Reddit, Product Hunt, and Indie Hackers, as described in Phases 1 and 2 of SKILL.md. * Boundary markers: The instructions in SKILL.md do not define specific delimiters or warnings to treat external web content as untrusted. * Capability inventory: The skill uses Write and Edit tools (defined in SKILL.md frontmatter) to store research findings and plans in local files such as market-opportunities.md and validation-plan.md. * Sanitization: No explicit sanitization or validation steps are defined in SKILL.md for the content retrieved from external searches before it is synthesized into actionable business plans.
Audit Metadata