paper-writing
Pass
Audited by Gen Agent Trust Hub on Aug 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process technical manuscripts, research evidence, and example paper corpora provided by the user. These represent ingestion points for untrusted data that could potentially contain hidden instructions to override agent behavior.
- Ingestion points: User-supplied paper drafts, example papers from local directories, and claim-evidence maps (as described in
SKILL.mdandreferences/group-paper-conventions.md). - Boundary markers: The skill lacks explicit boundary markers or delimiters (e.g., XML tags or unique markers) to separate user content from system instructions during interpolation.
- Capability inventory: The agent has the capability to read and write files within the project repository and execute shell commands for LaTeX compilation (
latexmk) as outlined inreferences/latex-project-conventions.md. - Sanitization: There is no mention of sanitization, filtering, or escaping for the content ingested from the external manuscript or example files.
Audit Metadata