accelerate-abilities-reference

name	accelerate-abilities-reference
description	What can this toolkit do? Full list of available capabilities and data queries. Advanced reference for power users.
license	MIT
category	reference
parent	accelerate

Accelerate — Capabilities reference

This is the internal reference for every Accelerate capability the toolkit can access. Use it when:

• The user asks "what can you do" or "give me the full list".
• Another skill needs to know whether a specific capability exists or what its inputs are.
• You need to pick the right capability for an unusual user question that doesn't fit a workflow skill.

Present the list to users in plain English, grouped by purpose. Do not show ability names, parameter names, or schemas unless the user explicitly asks for the technical detail (e.g., "I'm a developer, show me the raw list").

For the user — plain-English overview

When a user asks "what can you do", respond with something like:

I can help with four main things on your Accelerate site:

Understanding what's happening Weekly / monthly / real-time performance reports, top content, traffic sources, engagement metrics (bounce rate, time on page, scroll depth), trending content, concurrent visitors, underperformance diagnosis, content comparison across time periods, author performance, taxonomy breakdown.

Running A/B tests Propose test ideas grounded in your data, create tests with multiple variants, monitor live tests, declare winners, pause or stop tests, check experiment results with statistical confidence.

Personalising content Define audience segments (by referrer, UTM, location, device, behaviour), attach personalised versions of reusable blocks to audiences, update or review existing audiences.

Analysing campaigns and attribution UTM campaign performance, source breakdown, first-touch vs last-touch attribution, landing page analysis, channel comparison.

What do you want to dig into?

Offer to route them into the right specialised skill from there.

For the model — full technical reference

Every capability below is invoked via mcp__wordpress__mcp-adapter-execute-ability with the ability name and an input object matching the schema. Required inputs are bold.

Understanding what's happening (read-only)

• accelerate/get-performance-summary — site, block, post, or experiment metrics. Inputs: entity_type (site|block|post|experiment, default site), entity_id, date_range_preset (1h|4h|12h|24h|7d|30d|90d), date_range, granularity, segment.
• accelerate/get-post-performance — detailed metrics for one post. Inputs: post_id, date_range.
• accelerate/get-top-content — best-performing content. Inputs: type, limit (1–50, default 10), date_range.
• accelerate/get-content-diff — period-over-period comparison. Inputs: post_ids, current_period, comparison_period.
• accelerate/get-traffic-breakdown — breakdown by country / browser / os / referrer. Inputs: dimension (default country), date_range, limit.
• accelerate/get-taxonomy-performance — performance by category / tag / taxonomy. Inputs: taxonomy (default category), date_range, limit. Output includes underserved_score.
• accelerate/get-site-context — site design system (colours, typography, spacing, blocks). Inputs: include_blocks (default false).
• accelerate/get-engagement-metrics — bounce rate, time on page, scroll depth, recirculation, return visitor rate, exit pages. Inputs: entity_type (site|post), entity_id, date_range.
• accelerate/list-active-experiments — running tests and personalisation rules. Inputs: type (all|abtest|personalization), post_id.
• accelerate/list-experiments — historical and active experiments with pagination, status / type / date / annotation filters. Use this for "every experiment ever", not just live ones. Inputs: status (all|active|running|completed|paused|draft), type (all|abtest|personalization), date_range, subject_post_id, annotation_key, annotation_value, page, per_page (1–100, default 50). Returns paginated experiments with total and pages.
• accelerate/get-audience-segments — defined audiences. Inputs: include_estimates.
• accelerate/get-audience-fields — available targeting fields. Inputs: refresh.
• accelerate/get-author-performance — metrics for one or all authors. Inputs: author_id, date_range, limit, order_by (views|conversions|conversion_rate|posts_count).
• accelerate/get-author-content — posts by an author with metrics. Inputs: author_id, date_range, order_by, limit.
• accelerate/search-content — find content by title / URL / text. Inputs: query, search_in, post_type, limit.

Real-time

• accelerate/get-concurrent-visitors — live visitor count with polling guidance. Inputs: post_id, active_window_minutes (1–15, default 5), include_breakdown.
• accelerate/get-trending-content — content ranked by velocity (rising / flat / falling). Inputs: window (1h|4h|12h|24h, default 1h), metric (views|conversions|visitors), content_type, limit, min_threshold.

Attribution and campaigns

• accelerate/get-attribution-comparison — first-touch vs last-touch breakdown with insights (best introducer / closer / most undervalued). Inputs: date_range, conversion_goal.
• accelerate/get-source-breakdown — traffic by source / medium / campaign / referrer_domain. Inputs: date_range, group_by, limit.
• accelerate/get-utm-performance — UTM metrics by campaign / source / medium / term / content. Inputs: date_range, group_by, campaign.
• accelerate/get-landing-pages — top entry pages. Inputs: date_range, limit.

Running experiments (write, requires confirmation)

• accelerate/get-variants — inspect a block's current variants. Inputs: block_id. (Read-only.)
• accelerate/add-variant — add one variant to a block. Inputs: block_id, content, title, percentage, audience_id.
• accelerate/update-variant — modify an existing variant. Inputs: block_id, variant_index, content, title, percentage, audience_id.
• accelerate/remove-variant — delete a variant. Inputs: block_id, variant_index. Destructive.
• accelerate/create-ab-test — create a multi-variant test on a block. Inputs: block_id, variants (min 2), hypothesis, goal (engagement|click_any_link|submit_form), traffic_percentage.
• accelerate/set-block-goal — set a block's success metric. Inputs: block_id, goal.
• accelerate/set-traffic-percentage — adjust how much traffic enters the experiment. Inputs: block_id, percentage.
• accelerate/get-experiment-results — full statistical results with winner recommendation. Inputs: block_id, refresh.
• accelerate/stop-experiment — pause / resume / stop / declare winner. Inputs: block_id, action (pause|resume|stop|declare_winner), winner_variant_index (required for declare_winner).

Personalisation (write, requires confirmation)

• accelerate/create-audience — define an audience. Inputs: title, rules, description.
• accelerate/update-audience — modify an existing audience. Inputs: audience_id, title, description, rules.
• accelerate/create-personalization-rule — attach personalised content to a block for an audience. Inputs: block_id, audience_id, personalized_content, fallback_content.
• accelerate/broadcast-content — push blocks site-wide. Inputs: block_ids, title. Destructive.

Raw data (power users)

• accelerate/query-events — raw event queries with custom filters. Inputs: filters, fields, date_range, limit, offset.
• accelerate/get-event-schema — available event fields. Inputs: refresh.
• accelerate/aggregate — custom aggregations with flexible dimensions and metrics. Inputs: metrics, dimensions, filters, date_range, limit.
• accelerate/export-events — full event dump for a date. Inputs: date (YYYY-MM-DD), format (json|csv). Admin only.
• accelerate/get-export-status — check availability of exported data. Inputs: date.

Rules

• Always check capability existence before suggesting it. If a user asks about something Accelerate doesn't support (multivariate tests, revenue, heatmaps, etc.), say so directly.
• The reference above is exhaustive for the current toolkit. If a capability you expected isn't listed, it genuinely isn't available.
• Permission tiers: most "understanding" capabilities require can_view_analytics. Writing requires can_create_experiments. Broadcasts and exports require can_manage_experiments. If a call fails with a permission error, tell the user their WordPress account needs the right capability and suggest they check with their site admin.