doc-coauthoring

Pass

Audited by Gen Agent Trust Hub on Jul 25, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill implements a legitimate procedural workflow for collaborative writing, utilizing standard agent capabilities for file creation and modification.
  • [PROMPT_INJECTION]: The skill processes untrusted external data from shared documents, local files, and messaging platform history (Slack, Teams) to build documentation. This presents a surface for indirect prompt injection, as instructions embedded in these external sources could theoretically influence the agent's behavior during the co-authoring process.
  • Ingestion points: Shared documents via integrations, team channel history, and user-provided files or info-dumps.
  • Boundary markers: The skill does not explicitly instruct the agent to use delimiters or to disregard instructions found within the processed external content.
  • Capability inventory: The agent is granted capabilities to create and edit files (create_file, str_replace) and to invoke sub-agents for testing purposes.
  • Sanitization: No explicit sanitization or filtering of external data is defined in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 25, 2026, 08:03 PM
Security Audit — agent-trust-hub — doc-coauthoring