Forj

Create a challenge. Requires challenges:write and confirmation.

Create a competition. Confirm that this should be added to the authenticated gym. Requires competitions:write.

Create a competition event. Requires competitions:write and explicit confirmation.

Create a manual competition registration. Registration and payment-sensitive workflows require explicit confirmation.

Create an internal heat. Requires heats:write.

Create an internal heat event. Requires heats:write.

Create a programming assignment after explicit confirmation. Use only with validated Forj block JSON and the existing assignment scopes; never invent a parallel programming object.

Create a canonical Forj programming template. Requires one explicit confirmation and uses the same validation and schema as the Forj programming UI.

Create a draft week of programming assignments for the active gym. Accepts a structured weekly plan: for each day, supply the date, target scope (GYM, CLASS_TYPE, or GROUP), and the workout blocks using the existing Forj block schema. ALL assignments are saved as DRAFT status — they are not published or visible to members. Publishing is a separate Phase 3 action. Returns a summary: number of sessions created, movement variety, and any progression flags. Requires explicit confirmation before any data is written. Use list_programming_templates and list_programming_tracks first to pick valid targets. Requires OWNER or COACH role (admin:write scope).

Create a one-off scheduled action (future membership pause or reinstatement) from a previously previewed definition. Requires the confirmation_token from preview_scheduled_action. The stored definition is executed — it cannot be swapped.

Create a time-bounded trial or promotional membership offer. Preview shows: pricing, eligibility criteria (first-time only, per-member cap), validity window, and public visibility. Strong confirmation required due to financial and marketing impact. Backed by the existing membership plan routes (planType: TRIAL). Returns the created plan ID and a link to the member-facing sign-up page. Requires OWNER role (admin:write and billing:write scope).

Delete this competition and related operational data. This destructive change requires confirmation. Requires competitions:write.

Delete a competition event. Requires competitions:write and explicit confirmation.

Delete this heat. This destructive change requires confirmation. Requires heats:write.

Disable an automation rule so it stops running on its schedule. NOTE: pending executions already queued by past runs are NOT cancelled and will still run — tell the admin when any remain.

Generate draft message variants for a member outreach campaign WITHOUT resolving recipients or sending. Supply a cohort type (e.g. 'expiring_plans', 'inactive_members', 'new_members', 'overdue_payments', 'all_active', 'specific_members') and a campaign goal. Returns 2-3 message variants (subject + body) suitable for review before sending. No member data is accessed — this tool generates content only. Use preview_message_campaign next to resolve the actual recipient list with opt-out enforcement. Requires OWNER role (admin:write scope).

Enable a disabled automation rule so it runs on its schedule. Requires user confirmation. Before calling, run preview_automation_domain on the rule's domain and show the admin what currently matches.

Enroll a member in this challenge. Requires challenges:write and confirmation.

Dispatch an announcement campaign after the admin has confirmed the preview. Only call this after preview_announcement has returned a confirmation interaction and the admin has approved. Dispatches the campaign as an async background job — returns immediately with a job_id. Opt-out is re-enforced before each message is sent. The job can be cancelled within 60 seconds via cancel_job. Requires admin:write scope and an active gym context.

Apply a membership hold (pause) after the user has confirmed the preview. Only call this after preview_membership_hold has returned a confirmation interaction and the user has approved. Pass memberId, startDate, and endDate from the workflow's collected arguments. Requires support:write scope and an active gym context.

Verify how the currently deployed Forj product actually behaves when existing support knowledge is insufficient or the user explicitly needs implementation verification. First use support_search_features or support_get_feature for ordinary documented how-to questions; call this only when those results do not answer the behaviour question. Returns only a sanitised explanation, confidence, uncertainty, and escalation signal. It cannot return source code or accept repository, ref, branch, SHA, path, search, shell, or provider inputs. If status is insufficient_evidence, blocked, or unavailable, do not guess; explain the uncertainty and offer human support escalation.

Find live bookable drop-in class sessions at a gym for a given week. Returns sessions with start/end times, coach, capacity, number of spots booked, and drop-in pricing. Only sessions with public drop-in booking enabled are returned. Use book_class (with confirmation) to book one of the returned sessions.

Search the Forj platform for gyms available for membership or drop-in. Supports filtering by business type, city, suburb, postcode, or keyword search. Optionally supply caller GPS coordinates to sort results by proximity. Returns public gym profiles including slug, name, location, and booking availability. Never exposes internal gym IDs — use the returned slug with other tools.

Return the safe end-to-end workflow for generating Forj programming through MCP, including required reads, validation, confirmation and audit expectations.

Returns the caller's currently active gym context: gym name, slug, ID, and your role. When no gym is selected (multi-gym account), returns null for gym fields and lists all available gyms. Use this to diagnose 'no gym selected' errors or to check which gym is currently active before using gym-scoped tools. No scope required — any authenticated caller can invoke this tool.

Return members flagged as churn-risk with structured, explainable risk reasons. Each member is scored (low / medium / high) based on one or more of these signals: • attendance_gap — no check-in in the last 14 days. • failed_payment — has an overdue invoice. • plan_expiring_soon — active membership plan ends within 30 days. • benchmark_stale — has benchmark records but none logged in 60+ days. Risk score: high (3+ signals), medium (2 signals), low (1 signal). Results include memberId, displayName, riskScore, reasons array, and dataWindowDays. Use this for targeted outreach before members cancel. Requires member:read or retention:read scope.

Get details of a single automation action type, including its target model and payload expectations.

Get the full definition of one automation rule, including its domain, schedule, action payload, and run stats.

Get full details of one automation target model: field allowlist with types, allowed operators and known values, available actions, and the documented date-qualification rule. Use these exact field names when building domains.

Get challenge details and stats.

Read challenge leaderboard.

Retrieve the list of members assigned to the authenticated coach at their active gym. Only members explicitly assigned to this coach are returned — not all gym members. Requires COACH role or higher. Returns member profile details including name, status, and assignment counts.

Get a competition with its core configuration for the authenticated gym.

Read competition awards/placing summaries. Tenant is derived from auth.gymId.

Read the competition leaderboard. Tenant is derived from auth.gymId.

Surface readiness blockers and status for an upcoming competition. Checks: registration count vs expected, un-seeded divisions, events missing heats, judges not assigned, waiver completion rate, and payment completion. Returns a structured readiness report with a per-area status (ready/warning/blocked). Use this to identify what needs to be resolved before competition day. Then use preview_heat_publication to draft heats and publish_competition_heats to commit. Read-only. No confirmation required. Requires competitions:read scope.

Read sport-specific standings for a competition event (weightlifting or bracket standings).

Generate a structured weekly operating briefing for the active gym. Composes: attendance summary (sessions run, member visits, attendance rate, average class fill rate), billing health (unpaid invoices count, overdue amount, failed payment count), active programming week (tracks with content, any missing sessions), and upcoming competition blockers (competitions in the next 30 days with readiness status). Scoped to the active gym. Owner and admin roles only. Read-only. No confirmation required. Use this at the start of a ChatGPT session to understand the gym's current state before launching into any playbook (retention outreach, schedule optimisation, etc.). Requires admin:read scope.

Retrieve the public profile for a specific gym by its slug. Returns the gym's name, location, contact details, and website content. Only public-facing fields are returned — no internal billing, staff data, or configuration.

Returns a summary of the active gym including its name, slug, timezone, member count, and today's class schedule count. Requires an active gym context — call switch_active_gym first if no gym is selected. No specific scope required beyond authentication with an active gym.