doc-coauthoring
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process information from untrusted external sources, which creates a surface for indirect prompt injection attacks.
- Ingestion points: The skill explicitly instructs the agent to pull context from external integrations such as Slack, Microsoft Teams, Google Drive, and SharePoint, as well as user-provided files and links (documented in the 'Info Dumping' and 'Stage 1' sections of SKILL.md).
- Boundary markers: The instructions lack specific guidance for the agent to use boundary delimiters or 'ignore embedded instructions' warnings when processing this external data.
- Capability inventory: The skill utilizes powerful capabilities, including the ability to create files (
create_file), modify code or documents (str_replace), and invoke sub-agents for 'Reader Testing' (documented in the 'Stage 2' and 'Stage 3' sections). - Sanitization: No explicit sanitization, filtering, or schema validation is mentioned for the content retrieved from external sources before it is interpolated into the drafting and refinement process.
Audit Metadata