anthropic-doc-coauthoring

Pass

Audited by Gen Agent Trust Hub on May 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides a legitimate procedural workflow for document authoring and refinement. It acts as a guide for the user and does not contain any malicious patterns or hidden instructions.
  • [PROMPT_INJECTION]: While the skill is designed to ingest data from potentially untrusted external sources (such as team chat threads, shared documents, and user info-dumps), it does so within the context of helping the user draft a document. It does not attempt to bypass agent safety guidelines or override system instructions for malicious ends.
  • [COMMAND_EXECUTION]: The skill leverages file manipulation tools like create_file and str_replace to manage the drafting and iterative editing process. These operations are transparent to the user and are used as intended for document creation and maintenance.
Audit Metadata
Risk Level
SAFE
Analyzed
May 16, 2026, 07:14 PM
Security Audit — agent-trust-hub — anthropic-doc-coauthoring