mcpField
M
moderation-trust-review

moderation-trust-review

github.com/sylphxaiOther147 installsUpdated today
Repository

About moderation-trust-review

A skill for product and trust teams that turns content moderation questions into reviewable workflows and policies. It distinguishes policy uncertainty, quality issues, abuse, illegal content, emergencies, and repeat offenders, with escalation and appeal considerations.

Key features

  1. 01Content, actor, and harm assessment
  2. 02Moderation and appeal flow mapping
  3. 03Queue policy and enforcement ladder
  4. 04Transparency copy and event schema

Use cases

  1. 01Designing content review and appeal processes
  2. 02Defining enforcement for different harm categories
  3. 03Planning moderator queues and audit records

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/sylphxai/skills --skill moderation-trust-review

After installing, type/moderation-trust-reviewin Claude Code to confirm it loaded.

Bundled filesagents/openai.yamlreferences/moderation-trust-patterns.md

Details

From the project README

Moderation Trust Review

Use this skill to turn a recurring product-operations problem into a concrete, reviewable artifact instead of generic advice.

Workflow

  1. Identify content types, actors, harm classes, visibility states, enforcement authority, and appeal ownership.
  2. Read references/moderation-trust-patterns.md.
  3. Map submission, automated checks, queue triage, human review, enforcement, notification, appeal, and audit logging.
  4. Separate policy uncertainty, low-risk quality issues, abuse/fraud, illegal content, safety emergencies, and repeat offenders.
  5. Produce a moderation workflow, queue policy, enforcement ladder, transparency copy, and event schema.

When not to use

  • Do not use for generic advice the base model already handles without this skill's specific artifact contract.

Guardrails

  • Do not make irreversible account or creator decisions from low-confidence automation alone.
  • Do not hide enforcement reasons when a safe, specific explanation is possible.
  • Keep moderator tooling, user-facing policy, and product ranking logic separate.
  • Escalate legal, child safety, self-harm, or jurisdiction-specific obligations to qualified review.

Output format

Moderation context:
Content / actor / surface:

State flow:
- <state> -> <action / reviewer / user-visible effect>

Decision table:
| Scenario | Visibility | Enforcement | Appeal | Evidence |
| --- | --- | --- | --- | --- |

Queue policy and events:
- <queue/SLA/event>

Comments

Markdown supported