paper-revision
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from reviewer comments ($0) and the paper draft ($1).\n
- Ingestion points: Untrusted data enters the agent context through the reviewer feedback input and the paper draft files referenced in SKILL.md.\n
- Boundary markers: There are no explicit boundary markers or instructions to ignore embedded commands within the prompt templates in references/revision-prompts.md (e.g., using markers like [BEGIN REVIEWER COMMENT]).\n
- Capability inventory: The skill has the capability to modify local files (LaTeX source) and call other internal skills like experiment-code (which implies code execution), figure-generation, and self-review.\n
- Sanitization: There is no evidence of sanitization, validation, or escaping of the external reviewer feedback before it is interpolated into prompts for text generation or experiment design.
Audit Metadata