An agent can read site info, memory context and action history, add, correct, archive or delete persistent memory facts, run one of LolaCore's registered content or fleet actions, and read or edit posts and pages, all gated to administrators.
View LolaCore on WordPress.org
Needs WordPress 7.0 or later and the plugin active. Every ability here, named or bridged, checks manage_options, so only an administrator-level connected user can call any of them.
The 22 abilities listed are the fixed, literally-named registrations (14 curated abilities, 3 consumer content abilities, 5 LolaHub fleet abilities). Beyond these, a separate loop mirrors LolaCore's internal action registry (a code comment references 331 actions plugin-wide, including up to 92 injected by the Elementor addon) into lolacore/<action> abilities at runtime; those ids and their per-action readonly/destructive annotations depend on which addons are installed and licensed and which categories the owner has disabled, so they are not individually enumerated here. Three generic meta-abilities from the bundled WordPress MCP Adapter library (mcp-adapter/discover-abilities, mcp-adapter/get-ability-info, mcp-adapter/execute-ability) are also registered by LolaCore's bootstrap when no other active plugin has already registered them; they are omitted here because they are shared library abilities, not LolaCore-specific ones.
Grouped by what each one can change. Every one starts off until you enable it.
Returns a full site snapshot: WP/PHP versions, theme, active plugins, health, database size.
Retrieves persistent memory facts relevant to a given prompt.
Exports all memory facts as JSON for backup or audit.
Returns the history of actions executed on the site by any agent, filterable by date.
Returns the latest algorithmic site scan findings for SEO, security and performance.
Returns Lola's identity, memory, skills and playbooks in one call.
Lists and reads site content (posts, pages, custom types) with filters.
Returns an aggregated fleet health snapshot: versions, addons, pending updates, maintenance and security posture.
Exports this site's owner memory facts for a LolaHub pull/adoption.
Saves a new fact to persistent memory.
Corrects an existing memory fact by id.
Soft-deletes a memory fact by id; excluded from recall but recoverable.
Stores a new behavioral skill in an inactive state pending owner review.
Sends a message to the LolaCore chat assistant and returns its response.
Creates a new post or page (always saved as a draft on web-chat connections).
Edits a post or page's title, content, excerpt, slug, or status.
Applies playbooks, skills and preferences pushed from a connected LolaHub master.
Configures the AI provider and model used for local Lola chat.
Merges a batch of fleet-authored memory facts into this site's memory, deduping by hash.
Permanently deletes a memory fact by id.
Bulk-removes decayed and archived memory rows.
Runs an arbitrary action from LolaCore's internal action registry by id.
No abilities match that search.
Four steps from two active plugins to a governed connection.
Both plugins need to be active on the same WordPress site. LolaCore registers its own abilities; this plugin discovers and governs them.
In Agent Abilities for MCP, open the Bridge tab (labeled "Other plugins" in the admin). Abilities from every registered plugin appear there, grouped by the plugin that registered them.
Every ability from LolaCore starts off. Turn on only the ones you want an agent to reach, or enable the whole set in one click.
Approve the connection as a WordPress account scoped to what you actually want the agent to do. Each call still re-checks that account’s capability before it runs.
Read more in how bridging works or the full documentation.
Yes. It registers 22 fixed abilities directly in the core Abilities API, plus a dynamic bridge that mirrors its own internal action registry (hundreds of actions, including ones injected by addons like its Elementor integration) as individually-invocable abilities, so the Bridge can expose any of it once an operator turns them on.
the dynamic action-bridge portion (lolacore/<action>) skips any category the site owner has disabled and any premium addon action without an active license (AbilitySurfaceGate). a short internal skip-list keeps action ids that already have a dedicated named ability (e.g. add_memory_fact) from being registered twice.
Needs WordPress 7.0 or later and the plugin active. Every ability here, named or bridged, checks manage_options, so only an administrator-level connected user can call any of them.
An agent can read OttoKit's SaaS connection status, access-control settings, system health (cron jobs, database table, webhook log counts), configured automation triggers, which supported integrations are active, and paginated webhook execution logs.
Once an operator turns on the relevant feature toggles, an agent can read a site's exposed settings, users, and post content and terms, and can ask the plugin's AI experiments to draft or suggest excerpts, titles, slugs, meta descriptions, editorial notes and updates, summaries, translations, classifications, type-ahead text, image alt text, generated images, and comment moderation or reply suggestions, though running the content-generation ones also needs an AI connector configured.
An AI agent can create a new AutomatorWP automation, add triggers or actions to an existing one, and run or execute an automation's actions immediately, using AutomatorWP's own no-code automation builder.
Verified from LolaCore's own free wordpress.org code on 28 September 2026, not from its marketing claims. Not every ability listed here has been run end to end against a live site. Plugin authors change their code, so a name, a capability, or a condition can be different in a later release. This site is not affiliated with, or endorsed by, Fran Barbero or LolaCore. Back to the full directory.
Every ability starts off, is scoped to the user you connect, and is written to the same audit log as core. Turn on only what you need.
Off by default. Capability checked on every call.