Activity Nexus — PRD (Penny-proposed, Josh-greenlit)

Product: Objectuve (codename: Enkidu) Feature bundle: Connected Apps integration hub (Strava + Chess.com) with auto-mapping and idempotent check-ins Status: Shipped — built and merged to master 2026-08-03 (v4.14, narrative); not yet live to users — connected_apps_enabled flag at 0% pending Strava provisioning (B1–B3) Date: 2026-04-30 Proposer: Penny (Product Lead, feasibility study OBJ-44) PRD file: docs/product/completed/activity-nexus-prd.mdConfidence: Medium (concept strong, API landscape narrower than original ask)

1. Problem Statement

Users who already track fitness on Strava or hold streaks on Chess.com must re-enter the same data as Objectuve habit check-ins. Every redundant log-in is friction that breaks streaks and removes the people most likely to thrive — those already building activity habits elsewhere — from Objectuve's daily loop.

Specific gaps:

Strava users log a 10-mile run but must also check a "Morning Run" habit in Objectuve to count toward their goal. Two systems, two taps, one source of truth.
Chess.com players rack up games but don't link that activity to their "Play Chess" streak in Objectuve.
Duolingo streaks (the original ask) have no official API and present ToS risk — not feasible via public channels.
Fitbit is deprecated September 2026; the direct replacement is Google Health API, which Objectuve doesn't yet support.

The core opportunity: eliminate the double-entry by importing activity data and auto-checking the matching habits. This surfaces as a first-class "Connected Apps" surface, not a one-off integration.

2. Goals & Success Metrics

Primary Goals

Reduce friction for multi-tracker users — users who already log elsewhere should not have to re-enter data in Objectuve to maintain streaks.
Increase goal completion rate (north star) among users with active integrations by automating check-ins they would manually enter anyway.
Build integration framework for future providers — first with Strava + Chess.com (MVP), but designed reusable for Apple Health, Google Health, Garmin (post-partnership) without rewrites.
Demonstrate integration value for roadmap prioritization — if adoption is strong, integrations move from Year 3 to Phase 6 (concurrent with Teams monetization).

Success Metrics

Metric	Target	Measurement
% of users who connect at least one provider	8–15% (typical onboarding feature adoption)	Connected apps / active users
% of connected-app activities auto-checked	85%+	Auto-checked / total eligible activities (Strava runs, Chess.com games)
Avg goal completion rate (connected users vs. control)	+8–12pp	Cohort comparison post-launch
% of users who disconnect within 30 days	<20% (low churn = trust in mapping)	Disconnections / connections at 30d mark
Activity feed impression ("via Strava" marker)	50%+ of daily active users see it	Analytics on `activity_feed` events tagged `source: integration`
Strava OAuth completion rate	75%+ (Capacitor browser handoff)	OAuth grants / OAuth start attempts

Guardrails

Free tier feature — no paywall. Activity integration is free for all users. (See MISSION.md value #3, #5.)
Auto-check-in is confident, not silent — first connect surfaces explicit mapping review; subsequent new habits in matching categories are proposed (toast + link to settings), never silently checked in.
Credential security — encrypted token storage (Rails 7+ ActiveRecord::Encryption); tokens never logged or exposed in error messages.
ToS compliance — Strava GPS data requires Privacy Policy + ToS updates before any public launch. Internal/QA use allowed pre-update.
Idempotent check-ins — same activity can't double-check a habit. Check-in keyed on (goal_id, completed_date, integration_provider_id) uniqueness.
Graceful degradation — if a provider API is down, no check-ins are queued; users see a "Connection paused" message in settings, not an error on Dashboard.

3. User Stories

As a user with active Strava, I want to authorize Objectuve to read my activities, so that my fitness runs automatically check my habits without me manually re-entering.
As someone new to Connected Apps, I want to review a confidence-scored mapping of my Strava activities to my habits before confirming, so I trust the auto-check-ins and don't worry about mismatches.
As a user who built a Chess.com streak, I want my daily games to auto-check my "Play Chess" habit, so I maintain my Objectuve streak without extra work.
As a user concerned about privacy, I want to see exactly what data Objectuve imports and disconnect any time, so I control my integration footprint.
As a user whose integration breaks, I want a clear "Connection paused" message in settings, not a silent failure, so I can troubleshoot and reconnect.

4. Feasibility & API Landscape (Penny's Research)

Feasibility Score: 7 / 10

Dimension	Score	Notes
Technical complexity	6/10	OAuth flows + data normalization are real work; no integration framework exists yet in the Rails API
API availability	7/10	Strava and Chess.com are easy wins; Fitbit is deprecated Sept 2026; Duolingo has no official API
Market value	9/10	Integration hubs are proven retention drivers — reduces double-logging friction
North-star alignment	9/10	Auto check-ins → habits completed → goals completed. Direct.
Enkidu fit	7/10	Habits + recurrence are already there; need: integration framework, OAuth UI, auto-check-in logic
Roadmap timing	5/10	Integrations are Year 3; pulling forward is a deliberate choice

Supported Providers (MVP)

Provider	Status	Verdict	MVP Priority
Strava	✅ OAuth 2.0, free tier, well-documented	Webhook-driven, real-time activity ingestion	MVP
Chess.com	✅ Public REST API, no auth required	Daily polling, zero auth friction	MVP
Google Health API	✅ Active — correct Fitbit replacement	Post-MVP (framework supports it)	Phase 2
Apple HealthKit	✅ iOS native via Capacitor	Post-MVP	Phase 2
Fitbit	⚠️ Deprecated Sept 2026	Skip; use Google Health	Not supported
Garmin Connect	⚠️ Business partnership required	Post-MVP	Phase 3
Duolingo	❌ No official API, ToS risk	Do not build (use Chess.com instead)	Not supported

Critical finding (Penny): Duolingo has no official public API. Reverse-engineered community packages exist but violate Duolingo's ToS and are unsupported/fragile. Decision (Josh): Accept Chess.com as the streak-app substitute — fully public, zero auth, zero ToS risk.

5. MVP Scope (Josh's Direction)

What Ships in MVP

5.1 Strava Integration

OAuth 2.0 authorization (Capacitor in-app browser, callback via deep link)
Webhook subscription (real-time activity ingestion from Strava)
Activity normalization (Activity domain model with type, duration, distance, timestamp)
Mapping: Strava activity_type (run, cycle, swim, etc.) → Objectuve goal_category (fitness)

5.2 Chess.com Integration

Public REST API adapter (no auth)
Daily poller (Crono job at 02:00 UTC, fetches previous day's games)
Activity normalization (game count, timestamp)
Mapping: "player has games on date X" → Objectuve goal_category (games)

5.3 Smart Auto-Mapping Engine

Confidence-scored proposals on first connect (≥0.85 auto-check, 0.50–0.85 propose, <0.50 ignore)
Mapping review modal — user accepts, edits, or dismisses proposed matches
One-tap mapping management — Settings → Connected Apps → [Provider] → Manage Mappings
Toast notification on auto-map — "New habit detected: [habit name]. Linked via Strava. [Edit]"

5.4 Connected Apps Settings Surface

List of connected providers with status (connected, paused, error)
OAuth handoff for each provider
Disconnect (with confirmation)
View mapping details per provider
Re-run mapping review

5.5 Activity Feed Integration

Imported activities visibly attributed — "Checked in via Strava" pill on GoalEventCard
GoalEvent#source enum (manual | integration) to distinguish user-entered from imported
Analytics: track impression + tap-through rate on provider attribution

5.6 Documentation & ToS

User-facing docs: docs/features/connected-apps.md (how to connect, what data we access, how to disconnect)
Privacy Policy update (Termly): section on activity data import from Strava
ToS update (Termly): clause on third-party data ingest and retention
Release notes highlighting integration as free feature

What's Out of Scope (MVP)

Apple Health, Google Health, Garmin — Phase 2+ (framework designed for them, not built)
Bidirectional sync — Objectuve → Strava (sync check-ins back to Strava). One-way import only.
Goal progress updates — Only habits auto-check-in; goal progress (% toward monthly target) stays manual for now
Custom activity types — Only Strava's built-in types + Chess.com games mapped; no user-defined "other activity" type
Retroactive sync — Sync starts from connection date; historical data before connection date is not imported

6. Feature Design (High-level — UI-SPEC After Greenlight)

6.1 OAuth Flow (Strava)

Settings → Connected Apps → "Connect Strava"
Tap → Capacitor.Browser.open(strava_oauth_authorize_url) (in-app browser)
User logs into Strava, grants permission
Redirect to objectuve://strava/callback?code=...&state=...
Capacitor browser closes
Deep-link handler exchanges code for access token (backend POST /integrations/strava/authorize)
Mapping review modal surfaces (unless user opts to skip)
Token stored encrypted in IntegrationConnection record

Header: "Link your Strava activities to habits"
Proposed mappings table: (confidence score, activity type, → Objectuve habit, [Accept] [Edit] [Dismiss])
Help text: "High confidence (85%+) means we're confident; medium confidence means double-check this. You can change mappings anytime in settings."
Buttons: [Accept All] [Custom] (deep dive per mapping)
Confirmation state: "✓ 3 mappings linked. Your fitness activities will auto-check habits."

6.3 Connected Apps Settings Surface

List view:

Provider card per connected provider
Status badge (Connected / Paused / Error / Authorizing)
"Last sync: 2 hours ago" timestamp
[Disconnect] button (confirmation modal)
[Manage Mappings] link

Manage Mappings view:

Table: Strava activity type | Auto-check habit | [Edit] [Remove]
[+ Add Mapping] button
[Re-run Auto-Mapping] button (surfaces new mapping suggestions)

6.4 Activity Feed Attribution

Event card shows: "You checked in: Morning Run (via Strava) — 10 miles, 58 min"
Strava pill color: Strava orange (#fc5200), subtle background
Chess.com pill: Chess.com gradient, subtle background
Tap pill → attribution details (activity timestamp, synced data)

6.5 Disconnection & Error States

Disconnect: Confirm modal — "Remove Strava connection? Past check-ins stay, but future activities won't sync."
Paused state: "Your Strava connection paused. [Reconnect]" (user reconnects via OAuth, not manual token re-entry)
API error: "Strava sync failed. Check your connection and try [Retry]. Or [Disconnect]."

7. Technical Overview (Orion's Epic Plan)

7.1 Data Model

IntegrationProvider
  - provider_slug (Strava, Chess.com, GoogleHealth, etc.)
  - name, logo_url, description
  - auth_type (oauth, api_key, none)

IntegrationConnection
  - user_id, provider_id
  - access_token, refresh_token, token_expires_at (encrypted)
  - status (active, paused, error)
  - synced_at (last successful activity fetch)

HabitIntegrationMapping
  - user_id, integration_connection_id, goal_id (habit)
  - provider_activity_type (Strava.run, Chess.com.game, etc.)
  - confidence_score (0–1, used during proposal)
  - auto_check_enabled (boolean, user can disable per mapping)

IntegrationActivity
  - integration_connection_id, provider_activity_id, goal_id
  - activity_type, duration_minutes, distance_km, timestamp
  - checked_at (when auto-check-in happened, if any)
  - idempotency_key (provider_id, provider_activity_id, date) → prevents double-checks

7.2 Adapter Pattern

ruby

# lib/integration_provider/adapter.rb (abstract base)
class IntegrationProvider::Adapter
  def authorize_url(state:) # OAuth start → raises NotImplementedError
  def exchange_code(code:) # OAuth callback, returns tokens
  def fetch_recent_activity(connection:) # Returns array of normalized activities
  def verify_signature(request) # Webhook security (Strava specific)
end

# lib/integration_provider/strava.rb
class IntegrationProvider::Strava < IntegrationProvider::Adapter
  # Implementation of above methods
end

7.3 Interaction Layers

Integrations::AuthorizeConnection
  - Takes: provider_slug, oauth_code, user_id
  - Returns: IntegrationConnection record

Integrations::FetchAndMapActivities
  - Polls provider API or receives webhook
  - Calls Integrations::MapIncomingActivity for each activity
  - Calls Integrations::AutoCheckIn if confidence ≥ threshold

Integrations::MapIncomingActivity
  - AI-powered: analyzes activity description + user's habits
  - Returns: confidence score + proposed habit match

Integrations::AutoCheckIn
  - Wrapper around GoalTracking::CheckInHabit
  - Idempotent on (goal_id, completed_date, integration_provider_id)
  - Publishes domain event: integrations.activity_checked_in

7.4 Strava Webhook Receiver

POST /webhooks/strava
  - Signature verification (Strava HMAC-SHA256)
  - Parse activity subscription event
  - Enqueue `IngestionJob` (async)
  - Return 200 OK immediately

7.5 Chess.com Polling Job

ruby

# Crono schedule: 02:00 UTC daily
class IntegrationProvider::Chess::PollActivityJob
  - For each connected Chess.com user:
    - Fetch user's public games from "yesterday" (user's local timezone)
    - Dedupe vs. IntegrationActivity records
    - Call Integrations::FetchAndMapActivities

7.6 Confidence Scoring Algorithm

First-pass scoring (heuristic-based):

Rule 1: activity_type matches goal_category — graded by match quality, not binary:
- Exact match (activity_type maps 1:1 to the goal's category, e.g. Strava run/ride/swim/walk, or any Chess.com chess_<time_class> game) → 0.80 confidence
- Partial/family match (activity_type only loosely implies the category, e.g. a generic Strava workout) → 0.55 confidence
- No relationship → 0.0
Rule 2: Activity timestamp matches user's typical check-in time → +0.05
Rule 3: Goal has existing check-ins from same day → +0.05 (but capped at 0.95, avoids double-entry)
Rule 4: User has historically checked this habit on this day of week → +0.05

A bare exact category match with no other corroborating signal (e.g. a user's first-ever synced activity, before any check-in history exists) lands at 0.80 — surfaced for one-tap review, not silently auto-checked. This is intentional: "confident, not silent" (see A2 below) means the flagship first-connect review experience only fires once corroboration, not just a category match, earns auto-check.

Thresholds (tunable post-launch):

≥ 0.85 → Auto-check without user review
0.50–0.85 → Propose via toast notification + modal
< 0.50 → Ignore (don't surface or check)

Tuning: Post-launch, track false-positive rate (user edits / deletes auto-checks) and adjust thresholds in IntegrationConfiguration table.

8. Assumptions (Call Out If Wrong)

A1. Activity Nexus is a first-class product surface, not a bolt-on. The Phase 1 framework is reusable for future providers without architectural rewrites.
A2. Smart auto-mapping = confidence-scored proposal with accept/edit/dismiss, NOT silent zero-touch auto-mapping. First connect surfaces explicit review; new habits are proposed (toast), never silently checked.
A3. Strava uses webhooks for real-time activity ingestion. Chess.com uses daily poller. No new infra component needed; existing Sidekiq + Crono sufficient.
A4. Encrypted credential storage uses Rails 7+ ActiveRecord::Encryption (no third-party gem). Keys via Rails credentials / GCP Secret Manager.
A5. goal_category table will get a new games row (or similar) via seed migration to host Chess.com mappings. Fitness category already exists.
A6. Auto-check-ins reuse GoalTracking::CheckInHabit interaction (idempotent on (goal_id, completed_date)). We tag the resulting GoalEvent with source: integration so the feed can show "via Strava."
A7. User's time_zone is reliably populated (set in onboarding wizard or migrated from legacy system). If unset, Chess.com polling falls back to UTC with a TODO comment (precedent: GeneratePartnerMissedDayAlertsJob).

9. Open Questions (Resolve Before Phase 1 Dispatch)

Question	Resolution path	Blocker?
Strava rate limits (200 req/15min, 2000 req/day on free tier) — becomes constraint at ~5K connected users.	Apply for "increased rate limits" before public launch (tracked in Phase 6 release task).	No — document as A3 in release notes
Chess.com timezone handling — API returns UTC months; we compute "yesterday" in user's local TZ. Is `User#time_zone` always set?	Verify onboarding sets it; if not, add fallback (UTC + TODO). Phase 3 task includes guard.	No — has fallback
Confidence threshold tuning (currently 0.85/0.50/<0.50) — should these numbers be different?	Post-launch, monitor false-positive rate (user-deleted auto-checks) and adjust. Phase 4 task makes them constants.	No — has post-launch adjustment plan
iOS deep-link callback (Capacitor in-app browser closes on custom URL scheme).	Confirm `@capacitor/browser` handles `objectuve://strava/callback`. Risk: low (well-trodden pattern).	No — Phase 2 task includes verification
Strava redirect_uri for OAuth (sandbox vs. prod).	Use `ENV['STRAVA_REDIRECT_URI']` (different per environment). Phase 1 task includes configuration.	No — standard OAuth practice

10. Risks & Mitigations

Risk	Severity	Mitigation
Mis-mapped auto-check-ins erode trust	High	Confidence threshold + mandatory user confirmation on first connect. Never silent.
Webhook spoofing (Strava)	High	Strava signs webhooks; Phase 1 task includes signature verification. Reject unsigned events with 401.
Token leak via logs	High	Use Rails encrypted attributes AND filter `:access_token`, `:refresh_token` in logging config. Phase 1 includes this.
API credentials in Git	High	All secrets via `config/credentials.yml.enc` and GCP Secret Manager. Never commit plaintext.
Strava ToS breach (GPS data)	Medium	Phase 6 release task gates feature behind Privacy Policy + ToS update (Termly required).
Integration breaks post-launch, user confusion	Medium	Clear "Connection paused" status + [Reconnect] button. No silent failures.
Rate limit exceeded, integrations queue silently	Medium	Graceful degradation: failed fetches log warning + emit alert (Sentry), do not queue check-ins.
User deletes habit after mapping, integration tries to check non-existent goal	Low	`AutoCheckIn` validates goal exists; fails gracefully with warning log, does not fatal.
Timezone ambiguity (Chess.com)	Low	Fallback to UTC if `User#time_zone` unset. Documented in code comment.

11. Feature Design (PBC & Brand Check)

Free Tier Impact

✅ Extends free tier. Activity integration is fully free. No paywall, no gating to Supporter tier.

Engagement vs. Goals (North Star)

✅ Targets goals-completed (the north star) by automating check-ins users would manually enter anyway. Does NOT increase session length or notification volume. One-tap activity discovery, then users leave.

Dark-Pattern Check

✅ Confirmed clean. Auto-check-ins only happen for activities the user already created in another app (Strava run, Chess.com game). Not a phantom engagement hook. Attribution ("via Strava") is transparent.

Anti-Addictive Philosophy

✅ Aligns. Reduces the need to open Objectuve. Users with active integrations open less because progress happens automatically — the goal is fewer sessions, same or better outcomes.

12. Phase Breakdown (6 Phases, 6–10 Weeks AI Execution)

Phase	Title	Owner Mix	Outcome	Estimated size
53	Integration Framework Foundation	Codi (be) + Dori	`IntegrationProvider`, `IntegrationConnection`, `HabitIntegrationMapping`, `IntegrationActivity` models + adapter pattern + GraphQL surface + architecture doc	Large
54	Strava Integration (backend)	Codi (be)	OAuth flow, webhook receiver, `Strava::Adapter`, ingestion pipeline, 80% spec coverage	Large
55	Chess.com Integration (backend)	Codi (be)	Public REST adapter, Crono poller, daily aggregation, 80% spec coverage	Medium
56	Smart Auto-Mapping Engine	Codi (be)	`Integrations::MapIncomingActivity` interaction with confidence scoring, `Integrations::AutoCheckIn` (idempotent wrapper), `GoalEvent#source` enum + feed-marker glue	Large
57	Connected Apps Settings UI	Desi → Codi (fe)	UI-SPEC + Vue surface for connect/manage/map; activity feed marker; Storybook stories; Cypress smoke	Large
58	ToS, Privacy & Release	Dori + Codi (release)	Termly Privacy Policy + ToS updates; user-facing Connected Apps docs; CHANGELOG; feature-flag staged ramp	Medium

Total estimated AI execution time: 6–10 weeks (matching Penny's estimate for the epic once greenlit).

13. Test Strategy

Backend (RSpec + SimpleCov ≥80%)

Model validations/associations + uniqueness constraints
Interaction success + failure paths + side effects (job enqueuing)
GraphQL request specs for new queries + mutations
Webhook signature verification
Encrypted-attribute round-trip test
OAuth code → token exchange
Idempotency (same activity can't double-check)

Integration Tests (VCR cassettes)

Strava OAuth flow (with sandbox credentials)
Chess.com player profile + games-archive fetch
Webhook payload parsing + verification

Frontend (Vitest ≥80%)

Connected Apps list component (status badges, disconnect button)
Mapping review modal (confidence scores, accept/edit/dismiss)
Activity feed marker rendering (pill styling, attribution)
OAuth flow initiation (Capacitor browser open)

E2E / Smoke (Cypress)

Connect mock Strava provider, verify Connected Apps surface lists it
Disconnect, verify removal
(Real OAuth handoff stubbed via cy.intercept())

Manual QA (Pre-release)

Phase 57 mobile build with sandboxed Strava app
Known Chess.com username test account
Activity feed impression on mobile (pill rendering)

14. Dependencies & Ordering

Phase 53 must complete before 54, 55, 56 (foundation models + adapter pattern)
Phase 54 & 55 are parallelizable once 53 lands (independent adapters)
Phase 56 depends on 54 & 55 (MapIncomingActivity needs both adapters implemented)
Phase 57 depends on 53, 54, 55, 56 (UI surfaces framework outputs)
Phase 58 depends on 57 (needs working UI to document)

Suggested dispatch order: 53 (foundation) → 54 + 55 (parallel) → 56 → 57 → 58

15. Implementation Notes (For Codi During Phases 53–58)

Idempotency Strategy

Every imported activity must be idempotent to prevent double-checks. Use a composite key:

ruby

IntegrationActivity.find_or_create_by!(
  integration_connection_id: connection.id,
  provider_activity_id: strava_activity[:id],  # Strava's internal activity ID
  completed_date: activity_date
)

If the same Strava activity arrives twice (webhook retry, re-poll), the second find_or_create_by! returns the existing record — no second check-in.

Token Refresh

Strava tokens expire. On 401 response, refresh via Strava's refresh_token:

ruby

def refresh_access_token(connection)
  response = Strava::Client.post('/oauth/token', body: {
    client_id: ENV['STRAVA_CLIENT_ID'],
    client_secret: ENV['STRAVA_CLIENT_SECRET'],
    grant_type: 'refresh_token',
    refresh_token: connection.refresh_token
  })
  connection.update!(
    access_token: response[:access_token],
    refresh_token: response[:refresh_token],
    token_expires_at: response[:expires_at]
  )
end

Error Handling & Graceful Degradation

If a Strava webhook fails or a Chess.com poll times out, never queue a check-in. Instead:

Log the error to Sentry
Set connection.status = 'error'
Emit a Sentry alert
User sees "Connection paused. [Reconnect]" in settings

They must manually reconnect (re-do OAuth) to resume.

Staging & Production Credential Separation

yaml

# config/credentials/staging.yml.enc
strava:
  client_id: "staging-app-id"
  redirect_uri: "https://staging.objectuve.app/oauth/strava/callback"

# config/credentials/production.yml.enc
strava:
  client_id: "prod-app-id"
  redirect_uri: "https://objectuve.app/oauth/strava/callback"

Access in code: ENV['STRAVA_CLIENT_ID'] or Rails.application.credentials.strava.client_id

16. Rollback Strategy

All phases land behind a activity_nexus_enabled feature flag (reuses Feature Switches infra from item #37, already on roadmap). If needed:

Disable flag → No Connected Apps UI rendered, no webhooks processed (early-exit guard), no polling jobs run.
Existing manual check-ins unaffected (they use the standard GoalTracking::CheckInHabit interaction).
Rollback time: <5 minutes (flag flip via admin dashboard).

17. Suggested Routing (Post-Greenlight)

→ Orion to plan as Phase 53–58 milestone on the roadmap (epic, 6 phases).

→ Josh to confirm version number (originally proposed v1.12; align with current in-flight versions before committing).

→ Maggie to dispatch Phase 53 task package to Codi (foundation models first, specs/docs parallelizable).

→ Riley to bundle all 6 phases into one PR per phase on the milestone branch (gsd/v1.12-activity-nexus or equivalent) — phases merge sequentially as Vicki ships.

18. References & Evidence

Feasibility research: OBJ-44 comment by Penny (06b5aa97-e450-4469-822c-835d722d062c) — API landscape analysis, Duolingo ToS findings, Fitbit deprecation.
Epic plan: OBJ-44 comment by Orion (e2c17c7a-1000-4f32-8c50-4301a840b33f) — task breakdown, assumptions, risks, test strategy.
Josh's greenlight: OBJ-44 comment by Josh (772dc543-6ac9-4e91-b9ad-c26b5b935ff3) — Phase 6 timing, Chess.com substitute, smart auto-mapping, Strava ToS understood.
Strava API docs: https://developers.strava.com/docs/reference/
Chess.com API docs: https://www.chess.com/news/view/chess-com-api
Rails 7 ActiveRecord::Encryption: https://guides.rubyonrails.org/active_record_encryption.html
Capacitor Browser: https://capacitorjs.com/docs/apis/browser
MISSION.md values #3 ("Progress is Earned") and #5 ("Accountability, Not Entertainment") — anti-addictive alignment.

Appendix: FAQ

Q: Why not use Duolingo? A: No official public API. Community reverse-engineered packages violate ToS and are fragile. Chess.com is the safe substitute — fully public, no auth, strong market fit (chess learners overlap with habit-builders).

Q: Why is Fitbit not supported? A: Fitbit is deprecated September 2026. Google Health is the official replacement. It's on the roadmap for Phase 2, after MVP validates the integration framework.

Q: What if a user breaks an Objectuve habit and the integration tries to check it? A: AutoCheckIn validates goal exists. If goal is deleted, it fails gracefully (warning log, Sentry alert) and does NOT create a dangling event. This is tested in Phase 54 (Strava) and Phase 55 (Chess.com) spec suites.

Q: Can I export my data (Strava runs) from Objectuve later? A: This PRD does not include data export. One-way import only. Export is a separate feature proposal if needed post-launch.

Q: What about Apple Watch or Garmin devices? A: Apple Health is planned for Phase 2 (leverages Capacitor's existing HealthKit plugin). Garmin requires a business partnership; we'll apply if integration adoption is strong. See Phase 2+ roadmap.

Q: How long are activity records kept? A: IntegrationActivity records follow the same retention as GoalEvent records. Soft-delete (paranoia) applies. See docs/operations/data-retention.md for specifics (outside this PRD's scope).

Status: Ready for Epic Scoping

This PRD captures Penny's research, Orion's epic plan, and Josh's decisions. When Josh decides to begin Activity Nexus work, Maggie routes this PRD + OBJ-44 to Orion to invoke /gsd-new-milestone and start Phase 53 dispatch.

Last updated: 2026-08-02

Activity Nexus — PRD (Penny-proposed, Josh-greenlit) ​

1. Problem Statement ​

2. Goals & Success Metrics ​

Primary Goals ​

Success Metrics ​

Guardrails ​

3. User Stories ​

4. Feasibility & API Landscape (Penny's Research) ​

Feasibility Score: 7 / 10 ​

Supported Providers (MVP) ​

5. MVP Scope (Josh's Direction) ​

What Ships in MVP ​

5.1 Strava Integration ​

5.2 Chess.com Integration ​

5.3 Smart Auto-Mapping Engine ​

5.4 Connected Apps Settings Surface ​

5.5 Activity Feed Integration ​

5.6 Documentation & ToS ​

What's Out of Scope (MVP) ​

6. Feature Design (High-level — UI-SPEC After Greenlight) ​

6.1 OAuth Flow (Strava) ​

6.2 Mapping Review Modal (First Connect) ​

6.3 Connected Apps Settings Surface ​

6.4 Activity Feed Attribution ​

6.5 Disconnection & Error States ​

7. Technical Overview (Orion's Epic Plan) ​

7.1 Data Model ​

7.2 Adapter Pattern ​

7.3 Interaction Layers ​

7.4 Strava Webhook Receiver ​

7.5 Chess.com Polling Job ​

7.6 Confidence Scoring Algorithm ​

8. Assumptions (Call Out If Wrong) ​

9. Open Questions (Resolve Before Phase 1 Dispatch) ​

10. Risks & Mitigations ​

11. Feature Design (PBC & Brand Check) ​

Free Tier Impact ​

Engagement vs. Goals (North Star) ​

Dark-Pattern Check ​

Anti-Addictive Philosophy ​

12. Phase Breakdown (6 Phases, 6–10 Weeks AI Execution) ​

13. Test Strategy ​

Backend (RSpec + SimpleCov ≥80%) ​

Integration Tests (VCR cassettes) ​

Frontend (Vitest ≥80%) ​

E2E / Smoke (Cypress) ​

Manual QA (Pre-release) ​

14. Dependencies & Ordering ​

15. Implementation Notes (For Codi During Phases 53–58) ​

Idempotency Strategy ​

Token Refresh ​

Error Handling & Graceful Degradation ​

Staging & Production Credential Separation ​

16. Rollback Strategy ​

17. Suggested Routing (Post-Greenlight) ​

18. References & Evidence ​

Appendix: FAQ ​

Objectuve Docs

Activity Nexus — PRD (Penny-proposed, Josh-greenlit)

1. Problem Statement

2. Goals & Success Metrics

Primary Goals

Success Metrics

Guardrails

3. User Stories

4. Feasibility & API Landscape (Penny's Research)

Feasibility Score: 7 / 10

Supported Providers (MVP)

5. MVP Scope (Josh's Direction)

What Ships in MVP

5.1 Strava Integration

5.2 Chess.com Integration

5.3 Smart Auto-Mapping Engine

5.4 Connected Apps Settings Surface

5.5 Activity Feed Integration

5.6 Documentation & ToS

What's Out of Scope (MVP)

6. Feature Design (High-level — UI-SPEC After Greenlight)

6.1 OAuth Flow (Strava)

6.2 Mapping Review Modal (First Connect)

6.3 Connected Apps Settings Surface

6.4 Activity Feed Attribution

6.5 Disconnection & Error States

7. Technical Overview (Orion's Epic Plan)

7.1 Data Model

7.2 Adapter Pattern

7.3 Interaction Layers

7.4 Strava Webhook Receiver

7.5 Chess.com Polling Job

7.6 Confidence Scoring Algorithm

8. Assumptions (Call Out If Wrong)

9. Open Questions (Resolve Before Phase 1 Dispatch)

10. Risks & Mitigations

11. Feature Design (PBC & Brand Check)

Free Tier Impact

Engagement vs. Goals (North Star)

Dark-Pattern Check

Anti-Addictive Philosophy

12. Phase Breakdown (6 Phases, 6–10 Weeks AI Execution)

13. Test Strategy

Backend (RSpec + SimpleCov ≥80%)

Integration Tests (VCR cassettes)

Frontend (Vitest ≥80%)

E2E / Smoke (Cypress)

Manual QA (Pre-release)

14. Dependencies & Ordering

15. Implementation Notes (For Codi During Phases 53–58)

Idempotency Strategy

Token Refresh

Error Handling & Graceful Degradation

Staging & Production Credential Separation

16. Rollback Strategy

17. Suggested Routing (Post-Greenlight)

18. References & Evidence

Appendix: FAQ