present-paper
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFE
Full Analysis
- [DYNAMIC_EXECUTION]: The skill generates tailored Python scripts based on internal templates to inject speaker notes into presentations. This functionality is a core part of the skill's automation workflow and uses standard libraries.\n- [COMMAND_EXECUTION]: The skill uses Bash commands for file management, including unzipping PPTX archives for XML patching and converting PDFs to images. These operations are necessary for the technical manipulation of presentation files.\n- [INDIRECT_PROMPT_INJECTION]: The skill reads data from research papers (PDFs/DOIs) to generate summaries and scripts. This creates a surface for indirect prompt injection, but the risk is low as the content is extracted into a structured analytical framework.\n
- Ingestion points: Research paper content provided via PDF path, DOI, or PMID (SKILL.md).\n
- Boundary markers: None explicitly defined, though structured headers are used for the extraction phase.\n
- Capability inventory: File system access and script execution via Bash and Python-pptx library.\n
- Sanitization: Content is formatted into markdown templates without specific sanitization or filtering of external text.
Audit Metadata