api-docs-search
ChatGPTSearch the Ruddr public API documentation at docs.ruddr.io (REST API reference, OpenAPI endpoints, integration guides, webhooks, authentication). Use when the user asks how to call the Ruddr API, request/response shapes, endpoint URLs, query parameters, error codes, SDK or curl examples, or webhook payloads. Do not use for in-app product features or how-to questions about using the Ruddr web application — use kb-search for that. Returns the top matching pages, each with title, link, page path, and the page content (top of the page — full prose, including code examples and HTTP details). Cite results as clickable markdown links to the source page. Similarity search, no relevance threshold or score: an uncovered query still returns the closest pages. Judge each result against the question before citing it, and say the docs do not cover it when none apply.
kb-search
ChatGPTSearch Ruddr help documentation for feature explanations, configuration, and permissions. Returns the top matches, each as the section of the article that matched rather than the whole article: the page title, the headings that section sits under, a link straight to it, and its content. An article with no headings is returned whole. Cite a result as a markdown link to its link. If the section does not settle the question, search again with terms drawn from it — the rest of the article is not in this result. Similarity search, no relevance threshold or score: an uncovered query still returns the closest sections. Judge each result against the question before citing it, and say the docs do not cover it when none apply.
member-context
ChatGPTWho this connection acts as. Returns personal details of the connected member: their ID, name, security role, job title and time zone. Call it only when a request needs them, such as to resolve "me", "my" or "mine".
open-support-ticket
ChatGPTCreate a support ticket with the Ruddr support team on the user's behalf. Use when the user wants to submit a support ticket, report a bug, request a feature, or contact the Ruddr support team. First gather a clear subject and a detailed description, then confirm with the user before calling this tool. Do not call it speculatively or without explicit confirmation. Do not use for general product how-to questions that documentation can answer — use kb-search for those. On success the result includes the new ticket's id; tell the user their ticket was submitted and give them the ticket number for reference. Do not share the ticket link — the user cannot access it. If the call fails, direct them to email help@ruddr.io.
ruddr-command
ChatGPTChanges to Ruddr. Modes: search to find a command by namespace, load for its schema, run to execute; pass namespaces to search. A run is not idempotent: retrying one that may have succeeded can record a duplicate. track-time — recording time for the token's member: preflight, suggestions, and the entry itself
ruddr-query
ChatGPTRead-only Ruddr data. Modes: search to find a query by namespace, load for its schema, run to execute; pass namespaces to search. time — time tracking, time summaries, timesheet submissions, and submission coverage expenses — expense items, reports, and summaries clients — client records client-approvals — client approval requests for time and expenses projects — projects, tasks, project team members, project roles, and project health reports pipeline — companies, contacts, and opportunities billing — invoices, payments, credit notes, and receivables members — workspace members resourcing — resource allocations and staffing plans project-actuals — actual project hours, revenue, and financial performance project-plan — planned project hours, revenue, and financial performance project-forecast — forecasted project performance, revenue, and hours member-actuals — actual member hours, time off, revenue, performance, and utilization member-plan — planned member hours, revenue, performance, and utilization member-forecast — forecasted member hours, revenue, performance, and utilization project-budget — project overall budget configuration and budget vs actuals (revenue, performance, time) project-monthly-budget — project monthly budget configuration and budget vs actuals (revenue, time) custom-data — workspace taxonomies and lookup lists workspace — audit history
ruddr-status
ChatGPTCheck Ruddr's live service status and incident history from its public status page (status.ruddr.io). Returns the overall system status, per-component status (Public Website, Web Application, API, Mobile Apps, macOS Desktop App), recent incidents (active and recently resolved, with timestamps and impact), and any scheduled maintenance. Use when the user asks whether Ruddr is down or having problems, about outages or incidents now or on a given day ("Did Ruddr have an outage today?"), whether a component such as the API or web app is operational, uptime/availability, or upcoming maintenance windows. Takes no arguments — it always returns the full current status and recent incident history in one call. Do not use it for in-app product how-to or configuration (use the knowledge-base skill), the public REST API reference (use the ruddr-api-docs skill), or company/marketing questions (use the ruddr-website skill).
ruddr-website
ChatGPTFetch live content from Ruddr's public marketing website at ruddr.com (served by its Webflow CMS). Call with 'page' set to the page you need — a slug ('pricing'), a title ('Privacy Policy'), or a short topic ('about us'); the lookup resolves internally in a single call. If nothing matches, the site map comes back so you can retry with an exact slug. A result carrying match: 'fuzzy' was resolved by fuzzy search — verify its title answers the request before presenting it. Content ending in a '[Content truncated …]' marker is partial; link to the page URL for the rest. Call with no arguments only when you need the site map itself, e.g. to enumerate customer stories or list available pages. Use for questions about Ruddr the company and its public site: pricing and plans, feature overviews, founders and company background, customer stories, terms of service, and privacy policy. Do not use for in-app product features, configuration, or how-to questions — use kb-search. Do not use for the public REST API reference — use api-docs-search.
sitemap-directory
ChatGPTList the in-app navigation hierarchy — every page, record type, and report you can open — as a tree of nodes to choose from. Call this FIRST whenever the user wants to go to, open, or get a link to any part of the Ruddr app, then pass the chosen node id(s) to sitemap-url-builder. Each node carries an id (a URI: "project:tasks", "ready-to-bill", or "report:<slug>"). A node with an "entity" field can open a specific record (pass that record's name as entityQuery to the builder); a node with "params" accepts standard filters as query params; a list nests its record pages/tabs under "children". Only destinations the member is allowed to access are returned.
sitemap-url-builder
ChatGPTBuild the in-app navigation link(s) for node id(s) chosen from sitemap-directory. Call this once you have the node id(s) (from a directory fetched this turn or earlier this conversation). Pass the chosen nodeIds (a resource URI like "projects" / "project:tasks", or "report:<slug>"). To open a specific record, also pass entityQuery (the record's name/number) with a node that has an "entity". To deep-link, pass "params": the RQL where filters (e.g. { "where": { "project": { "billingType": ["tm"] } } }) plus navigation (period/start/end, date, year, unit, sort) — only params the node advertises are applied. For "my/me", filter members by the current member's name. Returns clickable link rows; render them for the user. Each row carries the app-relative 'url' and an 'absoluteUrl' with the site host — use absoluteUrl when writing a markdown link yourself. Never invent a link — only ids from the directory resolve, and an inaccessible one returns nothing.
skill-load
ChatGPTRead the instructions for one or more skills before acting on a matching request. A skill carries the rules for working with a part of this server — query grammar, write flows — that the tool schemas alone do not convey. Loading is free, has no side effects, and does not count against any limit. Use the names listed in the server instructions.
submit-feedback
ChatGPTSubmit the user's feedback about Ruddr to the Ruddr product team. Records the feedback, then notifies the team. Use when the user wants to send feedback, submit a feature request, suggest an improvement, or pass praise along to the Ruddr product team. First draft a concise subject and use the user's own words for the content, then present both to the user and get their explicit confirmation before calling this tool. Do not call it speculatively or without explicit confirmation — if the user has not reviewed and confirmed the subject and content, ask them to confirm first. Do not use for how-to or product questions that documentation can answer (use kb-search), for questions about the user's own workspace data (use the relevant data tool), or for support requests that need a response from the support team (use open-support-ticket). Examples (all already confirmed by the user): "Send feedback to the product team: please add dark mode — go ahead", "Submit this suggestion, I confirm: filter utilization by practice", "Let the team know the new approvals flow is great". Read the returned Status before replying. "submitted" means it was recorded and the team was notified — say so plainly. "saved_not_notified" means it was recorded but the notification did not go out, so say that the feedback is saved and the team will pick it up, and do not claim it was sent.
whats-new
ChatGPTworkspace-context
ChatGPTThe workspace this connection reads from: its ID, key, name, currency and time zone. The server instructions carry the same details as of connection time; call this when they may have changed.
api-docs-search
Claudekb-search
Claudeopen-support-ticket
Clauderuddr-command
Clauderuddr-context
Clauderuddr-query
Clauderuddr-status
Clauderuddr-website
Claudesitemap-directory
Claudesitemap-url-builder
Claudeskill-load
Claudesubmit-feedback
Claudewhats-new
Claude