Content Policing

Content Policing transforms WordPress content into an auditable graph of content, dynamic factual claims, entities, canonical facts, and evidence. It is designed for sites with hundreds or thousands of posts and pages where manually reviewing every article is impractical.

Claims are discovered dynamically from arbitrary prose. The plugin is not limited to predefined fact categories such as locations, prices, dates, or opening hours.

Features include:

• Incremental content fingerprinting so unchanged content can be reused.
• Structured claim extraction with a local deterministic extractor and an OpenAI-compatible schema-constrained provider.
• Entity resolution with aliases and claim/entity relationships.
• Internal conflict detection without assuming the most frequent value is correct.
• Flexible canonical facts with effective dates and temporal history.
• Evidence records with source, relevance, authority, freshness, and provenance metadata.
• Evidence-backed verification statuses including verified, contradicted, possibly stale, conflicting, unsupported, unverifiable, needs review, and not checked.
• Risk scoring and affected-content analysis when canonical facts change.
• Remediation previews, human approval, WordPress revisions, bulk review, and explicit apply confirmation.
• Asynchronous audits using Action Scheduler when available, with a WP-Cron fallback.
• A first-class, permission-aware MCP server with granular content, claim, fact, evidence, audit, conflict, and remediation tools.
• REST endpoints shared by the WordPress admin and domain services.
• Capability checks, nonces, rate limits, SSRF protections, audit logs, and secure token handling.

Does the plugin only check fixed fact types?

Does the plugin automatically rewrite published posts?

Can an LLM mark a claim as verified by itself?

Is MCP optional?

Does every audit run in one web request?

The optional OpenAI-compatible claim extraction provider sends selected post content to the administrator-configured endpoint. This is disabled unless configured by the administrator. Review the provider's terms and privacy policy before enabling it.

The optional evidence retrieval feature fetches only administrator-supplied HTTP or HTTPS URLs and applies URL, DNS, timeout, and response-size protections. It does not perform unrestricted web crawling.

The plugin stores claims, excerpts, entities, canonical facts, evidence metadata, audit records, remediation records, operational logs, and MCP request logs in the WordPress database. Configure retention and access controls according to your site's privacy requirements.