doc-coauthoring
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious code, obfuscation, or persistence mechanisms were found. The skill operates as described, assisting with document structure and refinement.
- [DATA_EXPOSURE]: The skill suggests using official platform integrations (e.g., Slack, Google Drive, Microsoft Teams) and MCP servers to pull context. This behavior is transparently described and gated by user settings and connector permissions.
- [INDIRECT_PROMPT_INJECTION]: The skill features a workflow that ingests external context to generate documents, which represents an indirect prompt injection surface.
- Ingestion points: Processes user-provided info dumps, content from linked files, and data retrieved via integrated messaging or storage tools in Stage 1.
- Boundary markers: No explicit delimiters or instructions are provided to the model to ignore embedded commands within the gathered context during the drafting phase.
- Capability inventory: The skill utilizes
create_file,str_replace, and sub-agent invocations (Stage 3) to process and finalize content. - Sanitization: No explicit sanitization or validation of the input context is performed before it is interpolated into drafting instructions.
Audit Metadata