doc-coauthoring
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns or security vulnerabilities were detected. The skill implements a standard professional workflow for document creation.
- [PROMPT_INJECTION]: The skill does not contain instructions to bypass safety filters or extract system prompts. It uses sub-agents for quality assurance as part of its documented 'Reader Testing' phase, which is a legitimate functional requirement.
- [DATA_EXFILTRATION]: No unauthorized data collection or transmission to unknown external domains. The skill utilizes official platform integrations (such as Slack, Teams, and Google Drive) for context gathering as intended by the user and within the platform's security model.
- [COMMAND_EXECUTION]: File operations are restricted to the local workspace for the purpose of drafting and editing the document using
create_fileandstr_replace. - [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it is designed to ingest data from external sources like shared documents and messaging channels. 1. Ingestion points: Shared documents and team chat history (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: File creation, surgical editing, and sub-agent invocation (SKILL.md). 4. Sanitization: Absent. The risk is mitigated by the structured, iterative workflow which requires human review of all drafted content before finalization.
Audit Metadata