papercut-script
Pass
Audited by Gen Agent Trust Hub on Oct 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from user-provided URLs. This creates a surface for indirect prompt injection where malicious instructions embedded in the source content could potentially influence the agent's behavior during the research and scripting process.
- Ingestion points: The skill uses web fetch and search tools to gather facts and UI details from external sources as described in the 'Research and rough UI' section of SKILL.md.
- Boundary markers: There are no specific instructions for the agent to use delimiters or 'ignore' tags when processing the external content.
- Capability inventory: The agent has the capability to perform web searches, fetch content, take browser screenshots, and write files to the local file system (e.g., narration/script.json, STORYBOARD.md).
- Sanitization: No explicit sanitization or validation steps are defined for the data retrieved from external links before it is used to generate the script and storyboard.
Audit Metadata