Skip to content
Uncategorized

WordPress AI Plugin 1.5.0: Practical Tools for Post Audio, Revision Clarity, and Image Provenance

WordPress’s official AI plugin 1.5.0 adds six opt-in experiments and closes two moderate security issues. Everything remains off by default.

WordPress’s official AI plugin reached 1.5.0 on 29 October 2026. The release adds six opt-in experiments and closes two moderate security issues. Everything remains off by default. You choose what runs.

The plugin continues its role as the Core AI team’s experimentation surface. Stable features stay supported. Experiments can change or disappear. Sites still need their own AI provider connector and API key.

What 1.5.0 Adds

  • Text to Speech — Generate audio for a post and render player controls on the frontend.
  • Revision Change Summary — Produce a plain-language description of differences between two revisions on both visual and classic revision screens.
  • Content Verification — Read-only detection of C2PA Content Credentials on uploaded JPEG, PNG, and WebP images.
  • MCP Access — Installs and activates the MCP Adapter plugin, then adds a Tools → MCP Access screen so you control which Abilities agents can reach.
  • WebMCP — Lets an agent browser operate inside the block editor via document.modelContext, with changes visible as they happen.
  • Block Container Label Generation — Suggests concise labels for container blocks in List View and the rename modal.

These sit alongside existing experiments such as excerpt generation, editorial updates, key encryption, and markdown feeds.

Security Fixes in This Release

  • GHSA-7mwf-4ppp-cxmp — Authenticated Author-level local file disclosure via an image path in Image Generation abilities.
  • GHSA-g94j-h78j-rr2j — Authenticated Contributor-level information disclosure via Content Classification.

Both are moderate. Update is recommended because the fixes are in the package and the official account urged prompt updates.

Why This Matters for Automation

The MCP Access and WebMCP experiments move WordPress closer to controlled agent workflows. Site owners can expose a defined set of Abilities and revoke access later. Revision summaries reduce the time spent reconstructing edit history, while C2PA detection provides a lightweight signal when images arrive with provenance metadata.

For teams already running Bridgistic or similar MCP bridges, the new MCP Access screen provides a single place to decide which capabilities agents may call. Text-to-speech creates an additional distribution surface without separate tools.

Practical Next Steps

  1. Update the AI plugin if it is already installed; it requires WordPress 7.0.3 or later.
  2. Review Settings → AI and enable only the experiments you intend to test on staging first.
  3. If you expose Abilities to external agents, use the new MCP Access screen to limit the set and log activity.
  4. Confirm your AI provider keys remain encrypted if you rely on the Key Encryption experiment.
  5. Check uploaded media for C2PA status only after enabling Content Verification; the check is read-only.

No automatic enablement occurs. The surface area grows only where you turn it on.

Leave a comment

Your email address will not be published. Required fields are marked *