paper-spine-latex
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external manuscript files, templates, and bibliographies, which represents a potential surface for indirect prompt injection if the inputs contain adversarial instructions.
- Ingestion points: Manuscripts, templates, and bibliography data are identified as primary inputs in SKILL.md and processed by the included Python scripts.
- Boundary markers: The instructions do not define specific tags or delimiters for manuscript content when it is interpolated into the agent's context.
- Capability inventory: The skill enables execution of LaTeX engines (pdflatex, bibtex), pandoc, and provided Python scripts for validation and conversion.
- Sanitization: The provided scripts (latex_guard.py, _paper_spine_utils.py) include normalization routines that strip comments and isolate prose from structural commands before performing checks.
Audit Metadata