coauthoring-documents

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The workflow relies on gathering and processing external context, including "existing materials" and "information dumps" provided by the user or fetched from the project environment. These ingestion points could serve as vectors for malicious instructions meant to influence the agent's drafting or testing behavior.
  • Ingestion points: Stage 1 encourages users to provide background context, history, and existing materials to reference (SKILL.md).
  • Boundary markers: The skill does not provide specific instructions or delimiters to isolate user-provided context from the agent's operational instructions.
  • Capability inventory: The agent is expected to draft documents and use tools like str_replace to modify files based on the processed context.
  • Sanitization: No sanitization or validation mechanisms are described for the external content before it is used in the refinement and drafting stages.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 08:32 AM
Security Audit — agent-trust-hub — coauthoring-documents