An agent can list the Orbit Fox modules on a WordPress site, see which ones are active and which settings can be changed, and activate or deactivate a module or change its settings.
Rooster One abilities, governed through the Bridge
An agent can read sales, cart, traffic, Search Console, customer and product data, write SEO titles, meta descriptions and image alt text, draft improved page body content, apply free SEO fixes, edit the local disavow list, and undo changes it applied.
- Abilities
- 7666 read10 write
- Active installs
- 10+
- Rating
- 100% (2 ratings)
- Version checked
- 11.3.9
- Requires
- WP 6.7+, PHP 8.1+
- Last updated
- 2026-11-09
- Abilities verified
- 10 November 2026
- Installs and rating as of
- 10 November 2026
View Rooster One on WordPress.org
What has to be true first
The Abilities API itself needs WordPress 6.9 or later (the plugin lists WordPress 6.7), plus PHP 8.1 or later and the plugin active. Of the 76 abilities, 59 check 'manage_woocommerce'; 3 check 'edit_pages'; 2 check 'edit_theme_options'; 1 check 'edit_posts'; for 11 the permission check runs through a helper or filter, so no single capability is listed here.
- Further abilities are registered through loops or helper code, so this list is a floor and their conditions were not checked.
This count can change
Counted from the plugin code on 10 November 2026; settings or feature flags that the count does not evaluate may hide some of these abilities.
The 76 Rooster One abilities
Grouped by what each one can change. Every one starts off until you enable it.
Read 66
- Read
Abandoned cart triage
Sessions with ≥1 cart-add and NO purchase-complete event, optionally filtered by minimum estimated cart value. Returns the pseudonym Customer #N, the top-5 products by quantity in that cart, the est_value, and last activity timestamp. Real names and emails NEVER leave the server.
- Read
Add-to-cart funnel by landing page
For the top-20 most-visited landing pages in the period, return the funnel: visitors → product viewers → cart-adders, plus cart-adds count. Joins wp_aieo_sessions_statics with wp_aieo_sessions and wp_aieo_addcart_events by session_id.
- Read
List Google Ads campaign plans
READ-ONLY, no AI credits. Lists the store's Google Ads campaign plans: name, status with a human label, objective (sales or leads), daily budget and currency, what the plan was built from, when it was created/pushed/activated, and a summary of its ad groups (headline and keyword counts, landing page). IMPORTANT for how you answer: a plan is created in Google Ads PAUSED and is NEVER auto-enabled, statuses are draft (nothing in Google yet), created (in Google, paused, spending nothing), activated (a person turned it on there), failed or archived. This ability only reports; it cannot compose, push, activate or archive a plan, and nothing it does can spend money. Params: status (optional filter), limit (default 50). Point the merchant to e-shop Manager → Ads to act.
- Read
Archive conversion rates
For
source=archivecart-adds, break down by archive term (pa_need / product_brand / product_cat) so the merchant can see which archives drive the most cart-adds. Parsessource_metaJSON. - Read
Attribution with engagement
Per attribution source (archive, single_product, sticky_atc, recos, blog_post, …) return cart-adds, distinct sessions, revenue, AND the engagement signals, avg session duration, avg scroll depth, avg time on the converting product view. Joins wp_aieo_addcart_events with wp_aieo_sessions / wp_aieo_sessions_statics by session_id.
- Read
Voice-of-customer survey summary
Summarise the post-order survey Voice-of-customer: NPS split (promoters/passives/detractors) and net score, VoC band and mean score, the flagged shipping-price-concern rate, mean satisfaction, the VoC→repurchase link (do promoters actually reorder more than detractors?), a count + preview of HIGH-VALUE detractors worth a recovery flow, and a sample of recent detractor open feedback. Read-only. Requires the Rooster One plugin with survey intelligence built.
- Read
Checkout friction (per-step timing)
Report where customers slow down or stall in checkout: the SLOWEST checkout step (from the per-step durations captured on the order-received page), overall active / idle / total-time percentiles, and a mobile-vs-desktop split. Surfaces friction the timing dashboard measures but never broke down per step. Read-only. Requires the Rooster One plugin; needs captured checkout-timing data.
- Read
Conversion-goal economic value
Value non-purchase conversion goals (leads, newsletter signups, course enrolments, form submissions) at their configured monetary worth over a window, so lead-generation campaigns are not scored as producing zero revenue. Returns per-goal completions × unit value, the grand total goal value, how many completions are unvalued (need a value set), and, optionally, the total goal value split by campaign (utm_campaign). Read-only. Requires the Rooster One plugin.
- Read
Blog engagement leaderboard
Top blog posts (and other non-product / non-archive pages) by total views, distinct visitors, avg time-on-page, and avg scroll-depth in the period. Sourced from wp_aieo_sessions page_view rows.
- Read
Blog → product conversion
For the top blog pages in the period, count visitors → product viewers (next click into a product) → cart-adders. Prefers source=blog_post attribution; falls back to same-session attribution if blog-post source rows are absent.
- Read
Read the boost slate
READ-ONLY, no AI credits. Reads the nightly BOOST FUSION slate, the stored ranking of what the shop should promote next across products, posts, pages, events and competitions. Per candidate: boost_score (0 to 100), the component signals behind it (SEO opportunity, search/commerce demand, sales forecast, on-site intent, waitlist demand, rating quality, timeliness, push appetite), the SEO quadrant, the social channels it is eligible for, and its VETOES. Vetoes come in two kinds and must never be reported as one: TRANSIENT (out_of_stock, stock_signout) means demand is real and only stock is missing, restocking clears it and the candidate returns to the top of the slate; PERMANENT (low_rating, ended) means the shop has decided not to amplify it at all. Params: limit (default 20), object_type (product|post|page|aieo_event|aieo_competition), status (all|promotable|blocked). Empty until the nightly rebuild has run. It never publishes, schedules or changes anything.
- Read
Restock to unlock, demand waiting on empty shelves
READ-ONLY, no AI credits. Lists the OUT-OF-STOCK products that customers have joined the waitlist for, with how many people are waiting on each, the total waiting, and the product's current boost score when the slate has one. This is the strongest demand signal in the shop because it was volunteered rather than inferred, a named person asked to be told when the product returns. Each of these carries a TRANSIENT veto on the boost slate: promotion is withheld only while the product cannot be bought, so restocking clears the block and turns the queue into a ready-made back-in-stock audience. Params: limit (default 25). Returns nothing when the waitlist feature is not in use. It never notifies anyone and never changes stock.
- Read
Brand gift coverage
For each active brand, report whether a Brand Gift offer is currently enabled (via
wp_aieo_dmm_brand_gift_offers) and which brands are uncovered. Use this to spot brands missing a gift hook. - Read
Brand performance tiers
Bucket every active brand into top / mid / bottom by lifetime revenue (from
wp_aieo_core_brand_vitals.total_sales). Top 20% = top tier, next 30% = mid, rest = bottom. Use this to triage the brand portfolio. - Read
Buy probability by engagement
For one product, return a 16-row bucket table (scroll-depth quartile × time-on-page quartile) with viewers, cart-adders, and empirical P(cart-add) per bucket. Heuristic, pure aggregate, not ML. Use to answer "if a visitor scrolls 75%+ and stays 90s+, how likely are they to buy?".
- Read
Buy-probability leaderboard
Products where cart-add rate among highly engaged visitors (scroll ≥75%, time on page ≥90s) is more than 5× the rate among disengaged visitors (scroll <25%, time <10s). Returns top N products by lift, the catalogue items where UX investment will pay off most.
- Read
Cart adds by source
Count add-to-cart events grouped by
source(archive, single_product, product_needs, sticky_atc, sidecart_*, …) for a period. The starting point for any attribution conversation. - Read
Cart-removal analysis
Products most often added then removed within the same session. Joins wp_aieo_addcart_events with wp_aieo_sessions remove_from_cart events by session_id × product_id. Returns top 50 by remove rate, buyers-remorse candidates.
- Read
DataForSEO connection status
Whether this site is connected to DataForSEO (SEO → General), the account login, current balance in USD and API rate limits. Free endpoint, costs nothing to call. Use before proposing any billed keyword/SERP data pull.
- Read
Dead archive detection
Archive pages (anchor_term_id) with reco impressions but ZERO cart-adds in the same period, the dead-zone candidates that may need a redesign, a hero, or to be retired.
- Read
Ranked keywords (DataForSEO snapshot)
Keywords this site ranks for on Google, from the weekly DataForSEO snapshot (free local read): keyword, position, search volume, ETV, intent, and the page. Filter to one object with object_type + object_id, or get the site-wide list by volume.
- Read
Striking-distance keywords
Keywords ranked 8 to 20 on Google (one push from page 1, the highest-ROI optimisation queue), highest volume first, with the page each one belongs to. Free local read of the weekly DataForSEO snapshot.
- Read
Content vitals leaderboard
Content (posts, pages, KB entries…) ranked from the daily content-vitals snapshot. order=demand (reach × engagement × recency), order=views (30-day views), or order=decaying (historically popular but fading, refresh candidates). Optional post_type filter.
- Read
Get cross-session entry credit
What each entry page earned INCLUDING customers who came back and bought days or weeks later, from
wp_aieo_entry_credit_daily(WooCommerce persisted attribution, dated by the order). Use this to value blog posts and other content that rarely sells in the same visit, where the same-session landing report shows zero. Group by content type to compare blog posts against products and categories. This is NOT the same measure as the landing report's purchases, which is same-session only, and the two will not add up. - Read
Find lookalike customers
Given a customer, return the top-N most similar customers by priority-inference distance (same feature vector the nightly k-NN job uses). All identifiers pseudonymous (
Customer #N). - Read
Find need gaps
Return
pa_needtaxonomy terms with fewer than the threshold number of products. These are the 'thin' needs whose archive pages won't drive conversion until more SKUs carry the tag. - Read
Find products needing ratings
Return products with many cart-adds in the period but zero or near-zero ratings, the "high commercial signal, no social proof" shortlist for rating-seeding campaigns.
- Read
Find a setting or feature
Answer "where do I set up / find X" AND "what does this button do" for anything in Rooster One. Searches every admin destination by label, id, description and live registry terms (courier carriers: ACS, BOX NOW, Courier Center, ELTA, Geniki, Speedex, DHL Express, DHL Paket, FedEx, UPS, USPS, Royal Mail, Colissimo, BRT, Correos Express, Sendcloud; marketplace feeds; order providers) and also every described operator ACTION. Results carry kind=page or kind=action. An action result tells you which screen it is on, whether pressing it emails the customer (informs_customer) and the confirmation question it asks. Use this before answering any "where is", "how do I enable" or "what happens if I click" question, and prefer it to guessing a URL. Never state that a control emails a customer unless informs_customer says so.
- Read
Find uncovered products
Return the highest-turnover products that have ZERO
pa_needtaxonomy terms assigned, the priority queue for needs-tagging work. Driven by Rooster One temp_orders, so only WC sites with the Rooster One order snapshot get rows. - Read
Form segments, who fills in a form
READ-ONLY, no AI credits. Turns a Gravity Form's choice fields (checkbox, radio, select, multiselect) into an audience breakdown: the ranked top segments that are useful for targeting, plus the full per-field distribution with counts and percentages. Answers "who is actually filling in my enquiry form?" and grounds ad copy, keywords and social targeting in what leads declared about themselves. Params: form (name, matched case-insensitively, partial names are fine) or form_id; limit (top segments, default 8); include_distributions (default true). Call it with NO form to list the forms available with their entry counts. Percentages are of RESPONDENTS to each field, not of entries, so on multi-select fields the shares sum past 100%. Returns nothing when Gravity Forms is not active. It never edits a form or an entry.
- Read
Get audience priorities
For one Rooster One audience key, return the inferred-priority distribution (which priorities matter most to that segment, by heuristic / kNN / embedding source). Pass audience_key=all for the whole base.
- Read
Get customer archetype
Return one customer's inferred priority profile (the F18-extrapolated archetype) plus supporting signals, price band, discount share, top brand affinity. Returns pseudonymised
Customer #Nonly; never real PII. - Read
Get customer profile
Return a customer's structured profile, inferred priority archetype, top brand affinities, wishlisted brands, rated brands, recent rating activity. Returns the pseudonym
Customer #Nonly; real names and emails NEVER leave the server. - Read
Get rating leaderboard
Return the top-N highest-rated and bottom-N lowest-rated products (with vote counts) from
wp_aieo_rating_stats.min_votesfilters out single-vote noise. - Read
Get reco conversion
Per-surface impression → cart-add conversion ratio. Joins
wp_aieo_reco_eventstowp_aieo_addcart_events(target_pid + surface alignment) to show which recommendation rails actually drive add-to-cart. - Read
Get reco impressions
Per-surface impression counts from
wp_aieo_reco_eventsfor a period. Use this to see which recommendation rails (same-needs, upsells, cross-sells, related, recently-viewed) are even being seen by shoppers. - Read
Landing pages by campaign
Traffic-source × landing-page performance from the canonical session stats (free local read): sessions, cart adders/adds, goal conversions per landing_path + utm_source/medium/campaign + referrer_host + channel over the last N days. Covers UNTAGGED referral traffic too: a plain link from another site shows as channel=referral with the referring host. Channels: paid_google (incl. gclid-only Google Ads), paid_facebook, social, organic_search, newsletter, referral, direct. Filter by utm_source, channel, or landing_path substring.
- Read
Exit pages
Where sessions END, from the canonical exit-page rollup (free local read): per exit_path over the last N days, exiting sessions, converted_sessions (the session fired a goal ANYWHERE before leaving, converters often keep browsing past the goal page), cart_sessions, average session duration. High exits with zero conversions = bleed points; high exits WITH conversions = natural end-of-journey pages.
- Read
List audiences
List the Rooster One audience segments configured on this store (guest, per-role, per-loyalty-tier) with sizes and labels. Requires the Rooster One plugin.
- Read
List brands
List all
product_brandtaxonomy terms with product counts, in-stock counts, and lifetime sales (fromwp_aieo_core_brand_vitals). - Read
List product needs
List every
pa_needtaxonomy term with the count of products carrying it. Use this to audit which needs are well-covered vs sparse. - Read
Need × conversion correlation
For each pa_need taxonomy term, aggregate cart-adds and product_views across all products tagged with that term and compute the term-level cart-add rate. Answers "which need terms drive the highest cart-add rate?".
- Read
Quote requests: demand a price list would have hidden
READ-ONLY, no AI credits. Lists the products visitors ASKED to be quoted, newest first, with how many people asked, how many are still unanswered, and which campaign sent them. This is volunteered demand: a named person typed their email to ask about a product the shop does not sell at a listed price, which is the only conversion a catalogue site can produce and never appears in any order report. Two counts are returned per product and MUST NOT be added together: "requested" is stated demand (they asked to be quoted) and "viewed" is browsing context (they looked at it while a quote was open). Params: days (default 30), limit (default 25), status (all|open|handled). Returns nothing when the product-requests feature is not in use. It never emails anyone, never changes a request and never exposes a quote token.
- Read
Rating × conversion correlation
Bucket products by their Rooster One average star rating (1-2, 2-3, 3-4, 4-5) and compute cart-add rate per bucket, SUM(cart_adds) / SUM(product_views), to answer "do my higher-rated products convert measurably better?". Joins wp_aieo_rating_stats with wp_aieo_addcart_events and wp_aieo_sessions.
- Read
Rating coverage by category
For each top-level
product_cat, return the number of products, the count with any votes, the average rating, and the vote density (votes / product). Use this to spot categories that need rating seeding. - Read
Recommend products for customer
Return a ranked list of product recommendations for one customer. Leverages the customer's wishlisted/affinity brands and the cached bundle recommendations (
wp_aieo_product_recommendations). PII-safe. - Read
Restock forecast
Which products need restocking. Returns items already out of stock (with the estimated lost-sales cost) plus items the self-tuning predictor projects will run out within
within_days, ranked by urgency. Each item includes the chosen daily sales rate, days until run-out, projected zero-stock date, a suggested reorder quantity, and the model's reliability_score / prediction_confidence (0 to 1) and uses_parent_signal flag. Choose grain="variation" for per-SKU restocking (default) or grain="parent" for whole-product planning. Reads wp_aieo_sales_predictor, rebuilt by the Product/Orders analytics "Sales / out-of-stock predictor" job. - Read
Get search trend
Search Console history from
wp_aieo_gsc_daily: the top search terms or pages for a window, each with the SAME window a year earlier, so seasonal movement can be told apart from real decline. Use this for "is this term growing", "what is seasonal" and "what did we rank for last summer". Returns nothing useful until Search Console has been connected long enough to hold a year, and says so rather than implying a collapse. - Read
List backlinks & toxic referring domains
READ-ONLY, no AI credits and no DataForSEO call, reads the latest stored backlink snapshot. Returns the profile summary (domain rank, backlinks, referring domains, spam score) and the most toxic referring domains (domain, spam_score 0-100, backlinks, whether already on the disavow list). Params: limit, min_spam (default = the toxic threshold). If the snapshot is empty, tell the merchant to open SEO → Backlinks and click Refresh (that call is metered on the DataForSEO budget).
- Read
List content gaps from customer chat
READ-ONLY, no AI credits. Returns the demand-weighted content gaps mined from the customer-facing chat: questions where the assistant found NO products (catalogue gaps) or the shopper rated the answer poorly (answer gaps). Each item has gap_key, a sample_question, occurrences, zero_product_turns, unhappy_turns, avg_rating, demand_weight, status, and (if an operator mapped it to a page) the mapped object + its live seo_score/open_fixes. Use it to find what content to CREATE or improve; then act with aieo/seo-apply-free-fixes (free) and the metered aieo/seo-fix-page / aieo/seo-fix-body. Params: limit, min_occurrences, only_unmapped, status (open|mapped|dismissed). Signals refresh on the AI Chat Analytics cadence.
- Read
List pages with open SEO fixes
READ-ONLY, no AI credits. Returns the highest-ROI pages that have open SEO fixes (fix_count > 0), ranked by real demand DESC then SEO score ASC, a high-demand page with a poor score sits first. Each item includes object_type, object_id, title, url, demand, seo_score, open_fixes, is_pillar, and has_pending_draft (whether an AI title/description draft is already awaiting approval). Also returns a store-wide summary (pages needing work, total open fixes, pillar count). Use this FIRST to answer "which pages need fixing" / "fix my top N pages" before calling aieo/seo-fix-page.
- Read
Google Search Console top queries / pages
READ-ONLY, no AI credits. Returns the site’s real Google organic performance from Search Console: rows of query (or page) with clicks, impressions, CTR and average position, over the last N days. Params: dimension (query|page), days (default 90), limit (default 100). Use it to see what actually brings traffic, spot high-impression low-CTR queries (a title/description fix opportunity), and striking-distance pages. If Search Console is not connected it returns connected=false with a hint to connect it on SEO → Search Queries.
- Read
Session journey for cart-add
Given a wp_aieo_addcart_events.id, return the ST session journey for that converting session: referrer, landing page, device, all prior product views in the session up to the cart-add, and the scroll-depth + time-on-page recorded for the converting product view.
- Read
Signal inventory (what this site measures)
The plugin's curated catalogue of the signals it computes, label, what it measures, grain, table, producing class/sproc, cron hook, consumers, enriched with live health: row counts, whether the producing cron is scheduled, and a state (healthy / empty / dormant / absent / unconsumed). Call this BEFORE proposing or building any new metric, scorer or rollup: most behavioural, order, customer, forecasting and SEO signals already exist here. Also the fastest way to answer "why is this report empty" (the answer is usually a dormant producer or a starved upstream) and "what are we computing that nobody reads". NOT exhaustive: it is a hand-maintained catalogue of the measurement estate, not a scan of every table, so a miss here is NOT proof the capability is absent, grep the tree before concluding anything has to be built.
- Read
List social boost candidates
READ-ONLY, no AI credits. Returns the top boost-worthy content objects for social promotion, ranked by the best available demand signal: the content-vitals content_demand score and the SEO store demand scores. Each item has object_type, object_id, title, score and url. Params: limit (default 10), type (post|page|product|any). If the signal tables are absent it falls back to recently published content. Use it to pick what to promote, then act with the social publishing abilities.
- Read
Social publishing performance
READ-ONLY, no AI credits. Per-platform organic social performance over a rolling window: posts_published and summed impressions, reach and link-clicks from the social post metrics, plus the most recent published posts. Params: days (default 30), platform (optional filter, e.g. facebook/instagram/tiktok). Use it to see which platforms are getting traction; it never publishes or changes anything.
- Read
Top converting reco pairs
Return the (anchor product, recommended product) pairs that convert best on a given surface. Filters by
min_impressionsso we don't surface noisy single-event 100% rates. - Read
List campaigns / workflows
List the store's marketing campaigns (workflows): name, trigger (order completed, abandoned checkout/cart, welcome, win-back, birthday, back-in-stock, opted-in/out, guest created, scheduled, manual), status (draft/active/paused), step count, audience, and each one's outcome funnel (sent, clicks, coupons redeemed, attributed revenue). Read-only. Requires the Rooster One plugin.
- Read
Get a campaign / workflow
Get one campaign (workflow) in full: its trigger and configuration, order/cart rules, steps (waits, sends with channel/offer, mutations), audience, and its outcome funnel. Read-only. Requires the Rooster One plugin.
- Read
Campaign performance
Report how the store's campaigns (workflows) performed: for each, the funnel sent → clicks (CTR) → coupons redeemed → attributed revenue, ranked by revenue, plus store-wide totals. Answers "how did my abandoned-cart / welcome / win-back campaign perform?". Pass an id for one campaign. Read-only. Requires the Rooster One plugin.
- Read
All marketing performance (owned + paid)
Report performance across ALL marketing in one place, owned automations (campaign workflows, always-on web-push triggers, broadcasts) AND paid ad campaigns, each with reach/sent, clicks, orders, revenue, spend and ROAS (paid) or revenue-per-1k-reached (owned), ranked by revenue, plus owned-vs-paid totals. Answers "what's working across all my marketing, per euro?". Read-only. Requires the Rooster One plugin.
- Read
Campaign incrementality (holdout lift)
For a campaign that reserves a HOLDOUT (a control slice deliberately not messaged), measure its TRUE incremental lift: treated-vs-held-out conversion rate and revenue-per-recipient, the percentage-point + relative lift, the incremental orders and revenue the send actually CAUSED, and a two-proportion z-test for significance. Answers "did this campaign really drive sales, or would those customers have bought anyway?". Read-only. Requires a workflow with a holdout that has run.
- Read
Campaign A/B test results
For a campaign whose message step runs an A/B copy test, report each variant's reach, open rate, click-through rate, conversions and reward, plus the current winner. The engine already leans traffic toward the winner (bandit). Read-only.
- Read
Multi-touch attribution
Credit owned marketing across the full purchase PATH, not just the last click: for each paid order, the messages (email / push) the customer engaged with beforehand share the revenue under a chosen model. Returns per-campaign credited revenue. Compare models (last_click / linear / time_decay / position) to see which campaigns last-click under-credits. Read-only.
- Read
Validate a page spec
Check a rooster-page@1 spec: the shape, the kinds, the keys, the icons, the patterns, the media, the links. Returns errors and warnings.
- Read
Export a page as a spec
A page's blocks back to a rooster-page@1 spec. Input: page (the recipe key, the slug or the id).
Write 10
- Write
Undo a change
Restore the state recorded when a change was applied. Input: undo_token.
- Write
Manage the disavow list & build the file
Edits the LOCAL disavow list only, it does NOT submit anything to Google (Google exposes no disavow API; the merchant uploads the generated file at the Search Console Disavow Links Tool). Actions (op): "add" a domain, "remove" a domain, "add_all_toxic" (every referring domain at/above the spam threshold), or "file" (just return the current disavow file). Always returns the up-to-date disavow file text + the tool URL + the list count. No AI credits. Requires the plugin admin capability.
- Write
Fix one page’s SEO title & description (AI)
GENERATIVE, CONSUMES AI CREDITS. For ONE page, writes an SEO title + meta description tuned to the live Google SERP for the page’s target keyword and its conversion goal, via the AI service. By default (apply=false) it DRAFTS the copy for the merchant to review (returns before/after); pass apply=true to publish it live immediately. Returns status ("drafted" | "applied" | "skipped"), the before/after title & description, and the target keyphrase. Because each call costs credits, confirm with the merchant before fixing more than a few pages. Get candidates from aieo/seo-list-fixes.
- Write
Apply free (no-AI) SEO fixes to a page
DETERMINISTIC, NO AI CREDITS. For ONE page, applies the structural fixes that need no copywriting: (1) canonical, materialise the canonical URL; (2) keyphrase, set the focus keyphrase from the page’s OWN top-ranked Google keyword when none is set. Applies immediately and re-scores, so the returned fix_count already reflects the change. Free and safe, always run this FIRST, before the metered generative fixes (aieo/seo-fix-page for title/description, aieo/seo-fix-image-alt for images). A THIRD fix, "indexable" (clear an explicit noindex so the page can rank), is EXPLICIT-ONLY: it is NEVER applied unless you pass fixes:["indexable"], because many pages are noindexed on purpose (thank-you pages, thin archives), only apply it after confirming the noindex was accidental. Returns the applied fixes + the new fix_count/score.
- Write
Write image alt text for a page (AI)
GENERATIVE, CONSUMES AI CREDITS. For ONE page, finds the images with NO alt text (featured image, product gallery, attached images) and writes a concise, descriptive, keyword-aware alt for each via the AI service, in the page’s language. By default (apply=false) it PREVIEWS the drafted alts; pass apply=true to write them to the media library. Returns per-image {id, filename, alt} and how many images were updated. Because it costs credits, confirm before running across many pages. Get candidates from aieo/seo-list-fixes.
- Write
Undo an auto-applied SEO fix
FREE. Reverts a focus keyphrase that was auto-set by aieo/seo-apply-free-fixes (it is source-stamped, so a manually edited keyphrase is never touched). Re-scores after reverting. Canonical is not undone because it is the correct value. Returns what was reverted.
- Write
Draft improved body content for a page (AI)
GENERATIVE, CONSUMES AI CREDITS. For ONE page, writes richer, better-structured body copy that naturally includes the focus keyphrase (clears the "keyphrase in content" check and fixes thin content), via the AI service, in the page’s language. SAFE BY DEFAULT: apply=false stores a DRAFT and returns a before/after excerpt, it does NOT change the live page. Pass apply=true to publish it to post_content; WordPress saves a revision automatically, so it can be rolled back from the page’s Revisions screen. Never invents facts. Confirm credit cost before doing several pages. Get candidates from aieo/seo-list-fixes or aieo/seo-content-gaps.
- Write
Generate social posts for a page (AI)
GENERATIVE, CONSUMES AI CREDITS. For ONE published object (product, post, page…), drafts platform-native social copy for the requested networks (facebook, instagram, x, linkedin, tiktok) via the AI service, in the page’s language. It only DRAFTS, nothing is published and no account tokens are used. Returns posts[], each with {platform, text, hashtags[], cta, variant_label}. Because it writes copy through the AI service, each call spends the merchant’s AI credits, confirm before generating across many pages.
- Write
Apply a page spec
Compile a rooster-page@1 spec into a page. mode draft (default) writes an autosave and returns a preview link (a new page is created as a draft); mode apply updates the page (a revision) and sets its meta. force: true, or the keys of the sections whose human edits to override.
- Write
Apply the menus
Write a rooster-menus@1 file's menus as navigation posts (adopted by slug). mode draft (default) plans without writing; apply writes.
No abilities match that search.
How to switch these on
Four steps from two active plugins to a governed connection.
Install Rooster One and Agent Abilities for MCP
Both plugins need to be active on the same WordPress site. Rooster One registers its own abilities; this plugin discovers and governs them.
Open the Plugins’ own tools screen
In Agent Abilities for MCP, open the Bridge tab (labeled "Plugins’ own tools" in the admin, "Other plugins" before 1.12.0). Abilities from every registered plugin appear there, grouped by the plugin that registered them.
Enable one ability at a time
Every ability from Rooster One starts off. Turn on only the ones you want an agent to reach, or enable the whole set in one click.
Connect as a least-privilege user
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.
Frequently asked questions
Does Rooster One work with the Bridge?
It depends on your setup. Rooster One calls wp_register_ability() for 76 abilities in its current code on wordpress.org, so the Bridge can discover them once the plugin is active. It also registers further abilities through loops or helper code that were not counted here.
What has to be true for Rooster One's abilities to register?
Further abilities are registered through loops or helper code, so this list is a floor and their conditions were not checked..
Does Rooster One need anything special to work with the Bridge?
The Abilities API itself needs WordPress 6.9 or later (the plugin lists WordPress 6.7), plus PHP 8.1 or later and the plugin active. Of the 76 abilities, 59 check 'manage_woocommerce'; 3 check 'edit_pages'; 2 check 'edit_theme_options'; 1 check 'edit_posts'; for 11 the permission check runs through a helper or filter, so no single capability is listed here.
More Other plugins
An agent can transliterate the existing non-Latin slugs of a scoped batch of posts and terms on a WordPress site into Latin characters. This is a write action that changes the stored slugs.
FluentBoards
ConditionalAn AI agent can list and read boards, stages, and tasks (including a personal task list and a cross-board search), then create, update, move, assign, archive, comment on and label tasks, and manage stages and labels.
Verified from Rooster One's own free wordpress.org code (version 11.3.9) on 10 November 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, Oxford Metadata Ltd or Rooster One. Back to the full directory.
Bridge Rooster One, keep the same control.
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.