mod-actions
A comprehensive moderation toolkit for Civitai, providing automated user management, strike systems, image review, content regulation, and CSAM reporting via tRPC API.
Introduction
The mod-actions skill is a powerful command-line toolkit designed for Civitai platform moderators to efficiently manage user behavior and platform content safety. By interfacing directly with the Civitai tRPC API, it allows administrators to execute complex moderation workflows with full side effects, ensuring platform integrity and policy enforcement. The tool is essential for community management, enabling rapid responses to ToS violations and maintaining high content standards through programmatic control.
-
User Moderation: Manage bans, mutes, leaderboard eligibility, and perform bulk content removal for bad actors using identifiers or usernames.
-
Strike Management: Execute a formal strike system including issuing points, tracking user standings, and voiding incorrect actions with detailed logs.
-
Image & Content Moderation: Process the review queue, flag TOS violations, handle ingestion errors, and manage model/training content approval or restriction.
-
Automated Safety Reporting: Dedicated module for NCMEC/CSAM reporting, allowing for efficient identification and escalation of illegal content.
-
Generation Regulation: Monitor flagged consumers, analyze generation restrictions, and perform audits on suspicious activity or restricted content patterns.
-
Usage requires an active Bearer token from a moderator-privileged Civitai account configured in a local .env file.
-
Commands are grouped by domain (query.mjs, strikes.mjs, images.mjs, etc.) and support both JSON output for integration pipelines and dry-run modes for safe verification.
-
Inputs typically require target IDs or usernames, while flaggable actions allow for optional reason codes like SexualMinor, Harassment, or BuzzCheating.
-
The tool operates in the production environment by default, though it can be redirected to staging instances via the CIVITAI_API_URL environment variable.
-
Ensure all destructive actions (such as ban or remove-content) are double-checked, as they perform permanent changes to the database state via the service layer.
Repository Stats
- Stars
- 7,101
- Forks
- 721
- Open Issues
- 103
- Language
- TypeScript
- Default Branch
- main
- Sync Status
- Idle
- Last Synced
- May 1, 2026, 07:14 AM