paper-ds-manage
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes project-specific data including a manifest file (paper-ds.json) and source code files (e.g., src/theme.css, src/components/*.tsx). This creates a potential surface for indirect prompt injection if these files contain malicious instructions.
- Ingestion points: Project manifest and component source files.
- Boundary markers: Not explicitly defined for separating code/data from instructions.
- Capability inventory: File system read/write and design canvas manipulation via sub-skills.
- Sanitization: Not present; the skill assumes the integrity of the local project files it manages.
- [DATA_EXFILTRATION]: The skill accesses project source code and design tokens to perform its primary function.
- Evidence: Accesses files within src/components and src/theme.css as defined in the manifest.
- Context: This access is restricted to the local project environment and is governed by explicit privacy rules in references/versioning.md that forbid publishing project history or assets without permission.
Audit Metadata