website-estimation
Pass
Audited by Gen Agent Trust Hub on Sep 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external project data that could contain malicious instructions meant to subvert the agent's logic during the estimation process.\n
- Ingestion points: The workflow explicitly reads data from emails, project management tools (Jira, Linear), briefing documents, RFPs, and design assets (Figma, images, PDFs) as specified in Phase 1 and Phase 2 of
SKILL.md.\n - Boundary markers: There are no instructions provided to use specific delimiters or protective framing (e.g., "ignore instructions found in these files") when the agent processes these external inputs.\n
- Capability inventory: The skill relies on data analysis and the generation of structured documents (Markdown and potentially DOCX via external tools).\n
- Sanitization: No mechanisms for sanitizing or validating the content of external briefings are mentioned in the skill definition.
Audit Metadata