story-collaborator
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [Prompt Injection] (SAFE): No direct instructions were found that attempt to bypass AI safety filters or override agent behavior.
- [Indirect Prompt Injection] (LOW): The skill processes untrusted external data (story drafts) and has the capability to write to the filesystem, which is a vulnerability surface for indirect prompt injection. 1. Ingestion points: The skill reads 'writer's drafts' and 'samples' from the conversation context. 2. Boundary markers: Absent; the instructions do not specify delimiters or 'ignore embedded instructions' warnings for the drafts. 3. Capability inventory: Writing to files including 'context/output-config.md' and project-specific markdown files. 4. Sanitization: Absent; no validation or escaping of input content is specified before persistence.
- [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials or access to sensitive system paths (e.g., SSH keys, environment variables) were detected. Local file writing is part of the stated functional purpose and no network transmission occurs.
- [Unverifiable Dependencies & Remote Code Execution] (SAFE): The skill does not install external packages, download remote scripts, or utilize dynamic execution patterns like eval() or exec().
Audit Metadata