recovery-community-moderator
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-generated content from recovery forums. While this represents a data ingestion surface where external actors could attempt prompt injection (e.g., trying to bypass moderation filters), the skill provides a robust Content Evaluation Framework with tiered severity (CRITICAL to PASS) and specific instructions to human-review ambiguous content. Boundary markers are recommended to prevent the agent from mistaking user input for instructions.
- [EXTERNAL_DOWNLOADS]: The skill references official resources from SAMHSA and the National Harm Reduction Coalition. These are well-known, trusted public health organizations and the references are purely informational.
- [COMMAND_EXECUTION]: The provided Python script
moderate_content.pyuses the official Anthropic API client to perform moderation tasks. It does not execute arbitrary shell commands or system utilities. - [DATA_EXPOSURE]: The skill uses structured JSON output for its moderation results, which is a best practice for safe data handling and prevents unintended data leakage in the UI.
Audit Metadata