lighting-report-extract
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from supplied PDF files and OCR results. Maliciously crafted text within these reports could potentially attempt to influence the agent's behavior.
- Ingestion points: The skill reads "supplied report PDF(s)" and uses "host PDF/OCR capabilities" to inspect text and table images (SKILL.md, Step 2).
- Boundary markers: The instructions tell the agent to preserve reported notation and avoid supplying values from memory (SKILL.md, Step 3), but there are no explicit delimiters or instructions to disregard potential commands embedded in the report text.
- Capability inventory: The skill has access to
Read,Write, andBashtools (SKILL.md, frontmatter). - Sanitization: No explicit validation or filtering logic is defined for the extracted text before it is processed or written to derived outputs.
Audit Metadata