← back to Cli Printing Press
chore(cli): remove stale planning docs (#56)
035d09741689653b8742e2162790a5b0137248c4 · 2026-03-29 14:52:22 -0700 · Trevin Chow
These research and audit docs from the cal.com CLI work are no longer
needed — the useful outcomes have been implemented.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Files touched
D docs/plans/2026-03-27-feat-calcom-cli-data-layer-spec.mdD docs/plans/2026-03-27-feat-calcom-cli-feature-parity-audit.mdD docs/plans/2026-03-27-feat-calcom-cli-power-user-workflows.mdD docs/plans/2026-03-27-feat-calcom-cli-research.mdD docs/plans/2026-03-27-feat-calcom-cli-visionary-research.mdD docs/plans/2026-03-27-fix-calcom-cli-audit.mdD docs/plans/dogfood-gauntlet-findings.mdD docs/plans/overnight-hardening-results.md
Diff
commit 035d09741689653b8742e2162790a5b0137248c4
Author: Trevin Chow <trevin@trevinchow.com>
Date: Sun Mar 29 14:52:22 2026 -0700
chore(cli): remove stale planning docs (#56)
These research and audit docs from the cal.com CLI work are no longer
needed — the useful outcomes have been implemented.
Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---
.../2026-03-27-feat-calcom-cli-data-layer-spec.md | 248 ---------------------
...6-03-27-feat-calcom-cli-feature-parity-audit.md | 101 ---------
...6-03-27-feat-calcom-cli-power-user-workflows.md | 124 -----------
docs/plans/2026-03-27-feat-calcom-cli-research.md | 56 -----
...026-03-27-feat-calcom-cli-visionary-research.md | 151 -------------
docs/plans/2026-03-27-fix-calcom-cli-audit.md | 45 ----
docs/plans/dogfood-gauntlet-findings.md | 74 ------
docs/plans/overnight-hardening-results.md | 78 -------
8 files changed, 877 deletions(-)
diff --git a/docs/plans/2026-03-27-feat-calcom-cli-data-layer-spec.md b/docs/plans/2026-03-27-feat-calcom-cli-data-layer-spec.md
deleted file mode 100644
index c3c32cb6..00000000
--- a/docs/plans/2026-03-27-feat-calcom-cli-data-layer-spec.md
+++ /dev/null
@@ -1,248 +0,0 @@
----
-title: "Data Layer Specification: Cal.com CLI"
-type: feat
-status: active
-date: 2026-03-27
-phase: "0.7"
-api: "cal.com"
----
-
-# Data Layer Specification: Cal.com CLI
-
-## Entity Classification
-
-| Entity | Type | Est. Volume | Update Freq | Temporal Field | Persistence Need |
-|--------|------|-------------|-------------|----------------|-----------------|
-| **Bookings** | Accumulating | 100-10,000/account | Daily | `createdAt`, `updatedAt` | SQLite + incremental sync |
-| **Event Types** | Reference | 5-50/account | Weekly | N/A | SQLite + periodic refresh |
-| **Attendees** | Accumulating (nested) | Same as bookings | With parent booking | via booking `updatedAt` | Extracted from booking data |
-| **Schedules** | Reference | 1-5/account | Monthly | N/A | SQLite + periodic refresh |
-| **Teams** | Reference | 1-10/account | Rarely | N/A | SQLite + periodic refresh |
-| **Calendars** | Ephemeral | 2-5 connections | Rarely | N/A | API-only (live check) |
-| **Webhooks** | Reference | 1-20/account | Rarely | N/A | API-only |
-| **Slots** | Ephemeral | Generated per query | N/A | N/A | API-only |
-
-## Data Gravity Scoring
-
-| Entity | Volume (0-3) | QueryFreq (0-3) | JoinDemand (0-2) | SearchNeed (0-2) | TemporalValue (0-2) | **Total** |
-|--------|-------------|-----------------|-----------------|-----------------|--------------------|-----------|
-| **Bookings** | 2 | 3 (daily: agenda, search) | 2 (event_types, attendees) | 2 (title, description) | 2 (trends, analytics) | **11** |
-| **Attendees** | 2 | 3 (daily: search by name/email) | 2 (linked to bookings) | 2 (name, email) | 1 | **10** |
-| **Event Types** | 1 | 2 (weekly: stale, stats) | 2 (referenced by bookings) | 1 (title, slug) | 0 | **6** |
-| **Schedules** | 0 | 1 | 1 | 0 | 0 | **2** |
-| **Teams** | 0 | 1 | 1 | 0 | 0 | **2** |
-
-**Primary entities (>= 8):** Bookings (11), Attendees (10)
-**Support entities (5-7):** Event Types (6)
-**API-only (< 5):** Schedules, Teams, Calendars, Webhooks, Slots
-
-## Social Signal Mining
-
-Evidence for local data persistence:
-1. No Cal.com backup/export tools exist (GitHub search confirmed)
-2. Calendar sync issues are #1 complaint — local copy is insurance
-3. callytics web dashboard (0 stars) tried analytics but no CLI
-4. n8n/Zapier integrations show demand for data out of Cal.com
-5. gcalcli (3k stars) proves calendar CLIs succeed by enabling offline workflows
-
-## SQLite Schema
-
-```sql
-CREATE TABLE IF NOT EXISTS bookings (
- id INTEGER PRIMARY KEY,
- uid TEXT NOT NULL UNIQUE,
- title TEXT NOT NULL DEFAULT '',
- description TEXT NOT NULL DEFAULT '',
- status TEXT NOT NULL DEFAULT '',
- start_time TEXT NOT NULL DEFAULT '',
- end_time TEXT NOT NULL DEFAULT '',
- duration INTEGER NOT NULL DEFAULT 0,
- location TEXT NOT NULL DEFAULT '',
- meeting_url TEXT NOT NULL DEFAULT '',
- event_type_id INTEGER NOT NULL DEFAULT 0,
- event_type_slug TEXT NOT NULL DEFAULT '',
- cancellation_reason TEXT NOT NULL DEFAULT '',
- rescheduling_reason TEXT NOT NULL DEFAULT '',
- rating REAL NOT NULL DEFAULT 0,
- absent_host INTEGER NOT NULL DEFAULT 0,
- created_at TEXT NOT NULL DEFAULT '',
- updated_at TEXT NOT NULL DEFAULT '',
- data TEXT NOT NULL DEFAULT '{}'
-);
-
-CREATE INDEX IF NOT EXISTS idx_bookings_status ON bookings(status);
-CREATE INDEX IF NOT EXISTS idx_bookings_start ON bookings(start_time);
-CREATE INDEX IF NOT EXISTS idx_bookings_updated ON bookings(updated_at);
-CREATE INDEX IF NOT EXISTS idx_bookings_event_type ON bookings(event_type_id);
-
-CREATE TABLE IF NOT EXISTS attendees (
- id INTEGER PRIMARY KEY AUTOINCREMENT,
- booking_id INTEGER NOT NULL REFERENCES bookings(id),
- booking_uid TEXT NOT NULL DEFAULT '',
- name TEXT NOT NULL DEFAULT '',
- email TEXT NOT NULL DEFAULT '',
- phone TEXT NOT NULL DEFAULT '',
- timezone TEXT NOT NULL DEFAULT '',
- absent INTEGER NOT NULL DEFAULT 0,
- language TEXT NOT NULL DEFAULT ''
-);
-
-CREATE INDEX IF NOT EXISTS idx_attendees_booking ON attendees(booking_id);
-CREATE INDEX IF NOT EXISTS idx_attendees_email ON attendees(email);
-
-CREATE TABLE IF NOT EXISTS event_types (
- id INTEGER PRIMARY KEY,
- title TEXT NOT NULL DEFAULT '',
- slug TEXT NOT NULL DEFAULT '',
- description TEXT NOT NULL DEFAULT '',
- length_minutes INTEGER NOT NULL DEFAULT 0,
- hidden INTEGER NOT NULL DEFAULT 0,
- price REAL NOT NULL DEFAULT 0,
- currency TEXT NOT NULL DEFAULT '',
- owner_id INTEGER NOT NULL DEFAULT 0,
- schedule_id INTEGER NOT NULL DEFAULT 0,
- booking_url TEXT NOT NULL DEFAULT '',
- data TEXT NOT NULL DEFAULT '{}'
-);
-
-CREATE INDEX IF NOT EXISTS idx_event_types_slug ON event_types(slug);
-CREATE INDEX IF NOT EXISTS idx_event_types_owner ON event_types(owner_id);
-
--- FTS5 on bookings
-CREATE VIRTUAL TABLE IF NOT EXISTS bookings_fts USING fts5(
- title, description,
- content='bookings', content_rowid='id'
-);
-
-CREATE TRIGGER IF NOT EXISTS bookings_ai AFTER INSERT ON bookings BEGIN
- INSERT INTO bookings_fts(rowid, title, description) VALUES (new.id, new.title, new.description);
-END;
-CREATE TRIGGER IF NOT EXISTS bookings_ad AFTER DELETE ON bookings BEGIN
- INSERT INTO bookings_fts(bookings_fts, rowid, title, description) VALUES ('delete', old.id, old.title, old.description);
-END;
-CREATE TRIGGER IF NOT EXISTS bookings_au AFTER UPDATE ON bookings BEGIN
- INSERT INTO bookings_fts(bookings_fts, rowid, title, description) VALUES ('delete', old.id, old.title, old.description);
- INSERT INTO bookings_fts(rowid, title, description) VALUES (new.id, new.title, new.description);
-END;
-
--- FTS5 on attendees
-CREATE VIRTUAL TABLE IF NOT EXISTS attendees_fts USING fts5(
- name, email,
- content='attendees', content_rowid='id'
-);
-
-CREATE TRIGGER IF NOT EXISTS attendees_ai AFTER INSERT ON attendees BEGIN
- INSERT INTO attendees_fts(rowid, name, email) VALUES (new.id, new.name, new.email);
-END;
-CREATE TRIGGER IF NOT EXISTS attendees_ad AFTER DELETE ON attendees BEGIN
- INSERT INTO attendees_fts(attendees_fts, rowid, name, email) VALUES ('delete', old.id, old.name, old.email);
-END;
-CREATE TRIGGER IF NOT EXISTS attendees_au AFTER UPDATE ON attendees BEGIN
- INSERT INTO attendees_fts(attendees_fts, rowid, name, email) VALUES ('delete', old.id, old.name, old.email);
- INSERT INTO attendees_fts(rowid, name, email) VALUES (new.id, new.name, new.email);
-END;
-
--- Sync state tracking
-CREATE TABLE IF NOT EXISTS sync_state (
- entity TEXT PRIMARY KEY,
- last_synced TEXT NOT NULL DEFAULT '',
- cursor TEXT NOT NULL DEFAULT '',
- total_items INTEGER NOT NULL DEFAULT 0
-);
-```
-
-## Sync Strategy
-
-### Bookings — Incremental via `afterUpdatedAt`
-
-**VALIDATED:** `afterUpdatedAt` query param confirmed in OpenAPI spec for GET /v2/bookings.
-
-1. Read `sync_state` for entity='bookings' to get last cursor
-2. GET /v2/bookings?afterUpdatedAt={cursor}&take=100&sortUpdatedAt=asc
-3. For each booking: upsert into `bookings`, delete+reinsert `attendees`
-4. Update sync_state with latest updatedAt
-5. Paginate with skip if hasNextPage
-6. Also sync cancelled separately (afterUpdatedAt + status=cancelled)
-
-**Batch size:** 100 (API max), **Rate budget:** 100 pages = ~50s at 120 req/min
-
-### Event Types — Full Refresh
-GET /v2/event-types → upsert all → update sync_state
-
-### Scoping Flags
-- `--team <id>` → adds teamId filter
-- `--since <date>` → overrides cursor
-- `--status <s>` → filters by status
-- `--max-pages N` → limits pagination (default: unlimited)
-
-## Search Filters
-
-| CLI Flag | SQL WHERE | Description |
-|----------|----------|-------------|
-| `--status` | `WHERE status = ?` | Filter by booking status |
-| `--attendee` | `JOIN attendees ... WHERE name LIKE ?` | Attendee name search |
-| `--email` | `JOIN attendees ... WHERE email = ?` | Exact email match |
-| `--event-type` | `WHERE event_type_id = ?` | Filter by event type |
-| `--after` | `WHERE start_time >= ?` | Start date |
-| `--before` | `WHERE start_time <= ?` | End date |
-| `--days N` | `WHERE start_time >= datetime('now', '-N days')` | Relative window |
-| (positional) | FTS5 MATCH on bookings_fts + attendees_fts | Free text search |
-
-## Compound Queries
-
-### 1. Search by attendee
-```sql
-SELECT b.*, a.name, a.email FROM bookings b
-JOIN attendees a ON b.id = a.booking_id
-JOIN attendees_fts ON attendees_fts.rowid = a.id
-WHERE attendees_fts MATCH ? ORDER BY b.start_time DESC LIMIT ?
-```
-
-### 2. Stats by event type
-```sql
-SELECT et.title, COUNT(*) as total,
- SUM(CASE WHEN b.status='cancelled' THEN 1 ELSE 0 END) as cancelled,
- ROUND(100.0*SUM(CASE WHEN b.status='past' THEN 1 ELSE 0 END)/COUNT(*),1) as show_rate
-FROM bookings b LEFT JOIN event_types et ON b.event_type_id = et.id
-WHERE b.start_time >= ? GROUP BY et.id ORDER BY total DESC
-```
-
-### 3. Stale event types
-```sql
-SELECT et.id, et.title, et.slug, COUNT(b.id) as cnt, MAX(b.start_time) as last
-FROM event_types et LEFT JOIN bookings b ON et.id = b.event_type_id
- AND b.start_time >= datetime('now', '-'||?||' days')
-GROUP BY et.id HAVING cnt = 0 ORDER BY et.title
-```
-
-### 4. Today's agenda
-```sql
-SELECT b.*, GROUP_CONCAT(a.name, ', ') as attendees
-FROM bookings b LEFT JOIN attendees a ON b.id = a.booking_id
-WHERE b.start_time >= ? AND b.start_time < ? AND b.status != 'cancelled'
-GROUP BY b.id ORDER BY b.start_time ASC
-```
-
-### 5. Conflicts (overlapping bookings)
-```sql
-SELECT b1.title, b1.start_time, b1.end_time, b2.title as conflict, b2.start_time as c_start
-FROM bookings b1 JOIN bookings b2 ON b1.id < b2.id
- AND b1.start_time < b2.end_time AND b1.end_time > b2.start_time
-WHERE b1.status != 'cancelled' AND b2.status != 'cancelled' AND b1.start_time >= ?
-```
-
-## Tail Strategy
-
-**Decision: REST polling** — Cal.com has no WebSocket/SSE for clients. Webhooks are push (server-to-server). Use afterUpdatedAt polling every 30-60s for a `tail` command.
-
-## Phase 4 Priority 0 Commands
-
-| Command | Data Path |
-|---------|-----------|
-| sync | WRITE: UpsertBooking + UpsertAttendees + UpsertEventType |
-| search | READ: bookings_fts + attendees_fts |
-| sql | READ: any table |
-| agenda | READ: bookings + attendees (compound query 4) |
-| stats | READ: bookings + event_types (compound query 2) |
-| stale | READ: event_types LEFT JOIN bookings (compound query 3) |
-| conflicts | READ: bookings self-join (compound query 5) |
diff --git a/docs/plans/2026-03-27-feat-calcom-cli-feature-parity-audit.md b/docs/plans/2026-03-27-feat-calcom-cli-feature-parity-audit.md
deleted file mode 100644
index c60b9ca2..00000000
--- a/docs/plans/2026-03-27-feat-calcom-cli-feature-parity-audit.md
+++ /dev/null
@@ -1,101 +0,0 @@
----
-title: "Feature Parity Audit: Cal.com CLI"
-type: feat
-status: active
-date: 2026-03-27
-phase: "0.6"
-api: "cal.com"
----
-
-# Feature Parity Audit: Cal.com CLI
-
-## Competitors Analyzed
-
-1. **@calcom/cli** (8 stars, TypeScript) — Official Cal.com CLI in companion repo. 46 command categories. Pure API wrapper.
-2. **gcalcli** (3k+ stars, Python) — Google Calendar CLI. Domain reference for calendar CLIs. 11 commands.
-
-## Feature Matrix
-
-### @calcom/cli (Official) — 46 Command Categories
-
-| Feature | @calcom/cli | gcalcli | Ours | Classification |
-|---------|-------------|---------|------|----------------|
-| **Booking CRUD** (list, get, create, cancel, reschedule) | YES | N/A (different API) | YES | TABLE STAKES |
-| **Event Type CRUD** (list, get, create, update, delete) | YES | N/A | YES | TABLE STAKES |
-| **Schedule CRUD** (list, get, create, update, delete) | YES | N/A | YES | TABLE STAKES |
-| **Calendar connections** (list, check, connect, disconnect) | YES | list calendars | YES | TABLE STAKES |
-| **Webhooks CRUD** | YES | N/A | YES | TABLE STAKES |
-| **Teams CRUD** | YES | N/A | YES | TABLE STAKES |
-| **Team event types** | YES | N/A | YES | NICE-TO-HAVE |
-| **Team memberships** | YES | N/A | YES | NICE-TO-HAVE |
-| **Conferencing apps** (list, connect, default) | YES | N/A | YES | TABLE STAKES |
-| **OAuth clients** | YES | N/A | NO | ANTI-SCOPE (platform feature, not user workflow) |
-| **Managed users** | YES | N/A | NO | ANTI-SCOPE (platform feature) |
-| **Organization-level operations** (131 endpoints) | YES | N/A | NO | ANTI-SCOPE (enterprise, org plan required) |
-| **Routing forms** | YES | N/A | NO | NICE-TO-HAVE |
-| **Verified resources** (email/phone verification) | YES | N/A | NO | ANTI-SCOPE (platform setup, not daily workflow) |
-| **Stripe integration** | YES | N/A | NO | NICE-TO-HAVE |
-| **Delegation credentials** | YES | N/A | NO | ANTI-SCOPE (enterprise only) |
-| **Login/auth** (OAuth PKCE flow) | YES | OAuth setup | YES | TABLE STAKES |
-| **Slots/availability query** | YES | N/A | YES | TABLE STAKES |
-| **Destination calendars** | YES | N/A | YES | NICE-TO-HAVE |
-| **Selected calendars** | YES | N/A | YES | NICE-TO-HAVE |
-| **Booking attendees** | YES | N/A | YES | TABLE STAKES |
-| **Booking confirm/decline** | YES | N/A | YES | TABLE STAKES |
-| **Private links** | YES | N/A | NO | NICE-TO-HAVE |
-| **Out-of-office** | YES (org only) | N/A | YES | NICE-TO-HAVE |
-| **Me/profile** | YES | N/A | YES | TABLE STAKES |
-
-### gcalcli (Domain Reference) — What Calendar CLI Users Expect
-
-| Feature | gcalcli | @calcom/cli | Ours | Classification |
-|---------|---------|-------------|------|----------------|
-| **Agenda view** (today's/week's events) | YES (core feature) | NO (just list) | YES | TABLE STAKES |
-| **Week calendar view** (calw) | YES | NO | NO | NICE-TO-HAVE |
-| **Month calendar view** (calm) | YES | NO | NO | NICE-TO-HAVE |
-| **Quick-add event** (natural language) | YES | NO | NO | NICE-TO-HAVE |
-| **Event reminders** | YES | NO | NO | NICE-TO-HAVE |
-| **ICS import** | YES | NO | NO | NICE-TO-HAVE |
-| **Updates since datetime** | YES | NO | YES (sync) | TABLE STAKES |
-| **Search events** | YES (text search) | NO | YES (FTS5) | TABLE STAKES |
-| **Colored output** | YES | NO | YES | TABLE STAKES |
-| **Configurable output format** | YES (--tsv, etc.) | JSON only | YES (--json, table) | TABLE STAKES |
-
-## Table Stakes Summary
-
-Features that MUST be in our CLI (from both competitors + domain expectations):
-
-| # | Feature | Source | Priority |
-|---|---------|--------|----------|
-| 1 | Booking CRUD (list, get, create, cancel, reschedule) | @calcom/cli | P1 |
-| 2 | Event Type CRUD (list, get, create, update, delete) | @calcom/cli | P1 |
-| 3 | Schedule CRUD (list, get, create, update, delete) | @calcom/cli | P1 |
-| 4 | Calendar connections (list, check) | @calcom/cli | P1 |
-| 5 | Agenda view (today's bookings, formatted) | gcalcli | P1 |
-| 6 | Search (full-text across bookings) | gcalcli | P1 |
-| 7 | Slot/availability query | @calcom/cli | P1 |
-| 8 | Conferencing apps (list, default) | @calcom/cli | P1 |
-| 9 | Webhooks CRUD | @calcom/cli | P1 |
-| 10 | Teams CRUD | @calcom/cli | P1 |
-| 11 | Me/profile | @calcom/cli | P1 |
-| 12 | Booking confirm/decline | @calcom/cli | P1 |
-| 13 | Colored/formatted output | gcalcli | P1 |
-| 14 | --json output mode | Standard | P1 |
-| 15 | Attendee management | @calcom/cli | P1 |
-
-## Anti-Scope Items (with cost analysis)
-
-| Feature | Why Anti-Scope | % Users Need | Competitor has it? |
-|---------|---------------|-------------|-------------------|
-| OAuth client management | Platform developer feature, not scheduling workflow | <5% | @calcom/cli (platform tier) |
-| Managed users | Platform feature for multi-tenant apps | <5% | @calcom/cli (platform tier) |
-| Organization-level ops (131 endpoints) | Requires org plan, enterprise-only | <10% | @calcom/cli |
-| Verified resources | One-time setup, not daily workflow | <5% | @calcom/cli |
-| Delegation credentials | Enterprise SSO feature | <2% | @calcom/cli |
-| Routing forms | Niche feature, web UI is better for form building | <10% | @calcom/cli |
-
-None of these anti-scope items have demand from users with >100 stars. The @calcom/cli has 8 stars total, so even its full feature set has no proven user demand. Our anti-scope decisions are safe.
-
-## Key Insight
-
-@calcom/cli has 46 command categories but is a pure API wrapper — no agenda, no search, no analytics, no health checks. gcalcli proves that calendar CLI users expect **agenda views and search**, not just CRUD. Our differentiator is combining @calcom/cli's API coverage with gcalcli's workflow intelligence, plus a local data layer that neither has.
diff --git a/docs/plans/2026-03-27-feat-calcom-cli-power-user-workflows.md b/docs/plans/2026-03-27-feat-calcom-cli-power-user-workflows.md
deleted file mode 100644
index 2a1a1cf9..00000000
--- a/docs/plans/2026-03-27-feat-calcom-cli-power-user-workflows.md
+++ /dev/null
@@ -1,124 +0,0 @@
----
-title: "Power User Workflows: Cal.com CLI"
-type: feat
-status: active
-date: 2026-03-27
-phase: "0.5"
-api: "cal.com"
----
-
-# Power User Workflows: Cal.com CLI
-
-## API Archetype: Scheduling Platform (Hybrid)
-
-Combines elements of:
-- **Project Management** — Event types = task types, bookings = tasks with states, teams with memberships
-- **CRM** — Attendees are contacts, booking pipeline tracks conversion from slot → booking → completion
-- **Infrastructure** — Calendar connections are integrations that break, conferencing apps need health checks
-
-## All 12 Workflow Ideas
-
-### 1. `agenda` — Today's Bookings Briefing
-- **What:** Show today's/this week's bookings with attendee info, times, event types, conferencing links
-- **API calls:** GET /v2/bookings?status=upcoming&afterStart=today&beforeEnd=tomorrow&sortStart=asc
-- **Frequency:** Daily (3) | **Pain:** High (3) — must open web UI | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 12/12**
-
-### 2. `search` — Full-Text Booking Search
-- **What:** FTS5 search across all synced bookings by attendee name/email, title, description
-- **API calls:** Sync bookings to SQLite → FTS5 MATCH query (no live API needed)
-- **Frequency:** Daily (3) | **Pain:** High (3) — no offline search exists | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 12/12**
-
-### 3. `stats` — Booking Analytics
-- **What:** Aggregate bookings: total, cancellation rate, show rate, most popular event types, busiest days/hours
-- **API calls:** Local DB query on synced bookings → aggregate by status/event type
-- **Frequency:** Weekly (2) | **Pain:** High (3) — no built-in analytics | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 11/12**
-
-### 4. `doctor` — Calendar + API Health Check
-- **What:** Check all calendar connections, conferencing apps, API auth, schedule validity
-- **API calls:** GET /v2/me → GET /v2/calendars/connections → GET /v2/calendars/{cal}/check → GET /v2/conferencing → GET /v2/schedules
-- **Frequency:** On-demand (2) | **Pain:** High (3) — sync issues are #1 complaint | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 11/12**
-
-### 5. `free` — Available Time Slots
-- **What:** Show when you're available for a specific event type over next N days
-- **API calls:** GET /v2/slots?username=X&eventTypeSlug=Y&start=A&end=B&timeZone=Z
-- **Frequency:** Daily (3) | **Pain:** High (3) — need to open app | **Feasibility:** Medium (2) — needs slot parsing | **Uniqueness:** 3
-- **Total: 11/12**
-
-### 6. `conflicts` — Double-Booking Detection
-- **What:** Find overlapping bookings across calendars, detect scheduling conflicts
-- **API calls:** GET /v2/calendars/busy-times + sync bookings → time range overlap detection in SQLite
-- **Frequency:** Weekly (2) | **Pain:** High (3) — silent double-books | **Feasibility:** Medium (2) | **Uniqueness:** 3
-- **Total: 10/12**
-
-### 7. `stale` — Unused Event Type Finder
-- **What:** Find event types with zero bookings in last N days, candidates for cleanup
-- **API calls:** Event types + bookings in local DB → cross-reference on eventTypeId
-- **Frequency:** Monthly (1) | **Pain:** Medium (2) | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 9/12**
-
-### 8. `unconfirmed` — Pending Booking Triage
-- **What:** List unconfirmed bookings, allow bulk confirm/decline
-- **API calls:** GET /v2/bookings?status=unconfirmed → POST confirm/decline
-- **Frequency:** Daily (3) | **Pain:** Medium (2) | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 11/12**
-
-### 9. `clone` — Event Type Cloning
-- **What:** Clone event type to another team or as personal copy
-- **API calls:** GET /v2/event-types/{id} → POST /v2/event-types or /v2/teams/{id}/event-types
-- **Frequency:** Monthly (1) | **Pain:** High (3) | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 10/12**
-
-### 10. `noshow` — No-Show Detection
-- **What:** Find past bookings not marked completed/absent, track no-show rate
-- **API calls:** Local DB → filter past unmarked → POST mark-absent
-- **Frequency:** Weekly (2) | **Pain:** Medium (2) | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 10/12**
-
-### 11. `export` — Booking Data Export
-- **What:** Export bookings to CSV/JSON for external analytics
-- **API calls:** Local DB → SELECT → format
-- **Frequency:** Monthly (1) | **Pain:** Medium (2) | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 9/12**
-
-### 12. `reassign` — Bulk Booking Reassignment
-- **What:** Move upcoming bookings from one host to another
-- **API calls:** GET bookings → filter by host → POST /v2/bookings/{uid}/reassign/{userId}
-- **Frequency:** Monthly (1) | **Pain:** High (3) | **Feasibility:** Easy (3) | **Uniqueness:** 3
-- **Total: 10/12**
-
-## Validation Against API Capabilities
-
-All 12 workflows validated against OpenAPI spec. Every required endpoint exists with the needed query parameters. Key validations:
-- `afterStart`, `beforeEnd`, `status`, `attendeeEmail`, `eventTypeId` filters on GET /v2/bookings ✓
-- GET /v2/slots supports username, eventTypeSlug, start, end, timeZone ✓
-- GET /v2/calendars/{cal}/check endpoint exists ✓
-- POST /v2/bookings/{uid}/confirm, /decline, /mark-absent, /reassign/{userId} all exist ✓
-- `afterUpdatedAt` supports incremental sync ✓
-
-## Top 7 for Implementation (Phase 4 Priority 2)
-
-| Priority | Name | Score | What It Does |
-|----------|------|-------|-------------|
-| 1 | **agenda** | 12 | Today's bookings with attendee info and times |
-| 2 | **search** | 12 | FTS5 full-text search across all bookings |
-| 3 | **stats** | 11 | Booking analytics: show rate, cancellation rate, busiest times |
-| 4 | **doctor** | 11 | Calendar connections, conferencing, API health check |
-| 5 | **free** | 11 | Available time slots for next N days |
-| 6 | **unconfirmed** | 11 | Pending bookings triage with bulk confirm/decline |
-| 7 | **conflicts** | 10 | Detect double-bookings and overlapping meetings |
-
-## Naming Pass
-
-| Command | Completes "I need to check ___" | Max 9 chars | No hyphens | No collision |
-|---------|--------------------------------|-------------|------------|-------------|
-| agenda | "my agenda" | 6 ✓ | ✓ | ✓ |
-| search | "for a meeting" | 6 ✓ | ✓ | ✓ |
-| stats | "my booking stats" | 5 ✓ | ✓ | ✓ |
-| doctor | "my calendar health" | 6 ✓ | ✓ | Standard |
-| free | "when I'm free" | 4 ✓ | ✓ | ✓ |
-| unconfirmed | "unconfirmed bookings" | 11 — over limit but clear | ✓ | ✓ |
-| conflicts | "for conflicts" | 9 ✓ | ✓ | ✓ |
diff --git a/docs/plans/2026-03-27-feat-calcom-cli-research.md b/docs/plans/2026-03-27-feat-calcom-cli-research.md
deleted file mode 100644
index 831d1d14..00000000
--- a/docs/plans/2026-03-27-feat-calcom-cli-research.md
+++ /dev/null
@@ -1,56 +0,0 @@
----
-title: "Research: Cal.com CLI"
-type: feat
-status: active
-date: 2026-03-27
-phase: "1"
-api: "cal.com"
----
-
-# Research: Cal.com CLI
-
-## Spec Discovery
-- **Official OpenAPI spec:** https://github.com/calcom/cal.com/blob/main/docs/api-reference/v2/openapi.json
-- **Source:** GitHub repo, confirmed via discussion #17565
-- **Format:** OpenAPI 3.0.0 JSON (1.1MB)
-- **Total paths:** 181 (103 user-level, 78 org-level)
-- **Total operations:** 285 (154 user-level)
-- **Resource categories:** bookings, event-types, schedules, calendars, conferencing, webhooks, teams, slots, me, oauth-clients, verified-resources, routing-forms, stripe, selected-calendars, destination-calendars, api-keys
-- **API version header:** `cal-api-version: 2026-02-25` (date-based, required)
-
-## Competitors (Deep Analysis)
-
-### @calcom/cli (8 stars)
-- **Repo:** https://github.com/calcom/companion (packages/cli/)
-- **Language:** TypeScript (98.6%)
-- **Commands:** 46 command categories (auto-generated from OpenAPI)
-- **Last commit:** February 2026
-- **Maintained:** Yes (active development)
-- **Notable features:** OAuth PKCE login, full API v2 coverage, openapi-ts code generation
-- **Weaknesses:** Pure API wrapper. No agenda view, no search, no analytics, no health checks. 8 stars despite Cal.com's 34k. Requires Node runtime.
-
-### gcalcli (3k+ stars) — Domain Reference
-- **Repo:** https://github.com/insanum/gcalcli
-- **Language:** Python
-- **Commands:** 11 (agenda, calw, calm, list, edit, add, quick, import, remind, updates, init)
-- **Maintained:** Yes
-- **Notable features:** Agenda view, week/month calendars, colored output, reminders, search
-- **Relevance:** Proves demand for calendar CLIs
-
-## User Pain Points
-> "Some API endpoints don't work properly... resorting to tRPC calls" — Issue #14706
-> "Workflows can only be created/edited through the UI" — Issue #25560
-> Calendar sync issues (Google #8376, Outlook #3546) are the #1 complaint
-
-## Auth Method
-- **Type:** Bearer token (API key prefix: `cal_live_`) or OAuth 2.0 PKCE
-- **Env var:** `CAL_COM_API_KEY`
-- **Rate limit:** 120 req/min
-
-## Strategic Justification
-@calcom/cli has 8 stars vs Cal.com's 34k — that's a vacuum, not competition. It's a pure API wrapper with no workflow intelligence. gcalcli (3k stars) proves calendar CLI demand. Our differentiators: SQLite data layer with FTS5, booking analytics, agenda view, Go single binary, agent-native flags.
-
-## Target
-- **Commands:** 40+ (15 CRUD + 7 workflows + 7 data layer + 11 supporting)
-- **Differentiator:** Local SQLite + FTS5 search + booking analytics + agenda
-- **Quality bar:** Grade A (80+/100)
diff --git a/docs/plans/2026-03-27-feat-calcom-cli-visionary-research.md b/docs/plans/2026-03-27-feat-calcom-cli-visionary-research.md
deleted file mode 100644
index 9465b4f9..00000000
--- a/docs/plans/2026-03-27-feat-calcom-cli-visionary-research.md
+++ /dev/null
@@ -1,151 +0,0 @@
----
-title: "Visionary Research: Cal.com CLI"
-type: feat
-status: active
-date: 2026-03-27
-phase: "0"
-api: "cal.com"
----
-
-# Visionary Research: Cal.com CLI
-
-## Overview
-
-Cal.com is open-source scheduling infrastructure ("the open-source Calendly alternative") with 34k+ GitHub stars. It serves businesses, freelancers, and developers who need scheduling automation. The API v2 has 181 paths / 285 operations covering bookings, event types, schedules, calendars, teams, organizations, webhooks, conferencing, slots, and more. The official @calcom/cli exists in the companion repo (8 stars, TypeScript, 46 command categories) but has near-zero adoption — it's an auto-generated API wrapper with no workflow intelligence. API v1 is deprecated and shuts down April 8, 2026 — making v2-native tooling timely.
-
-The strategic opportunity: build a CLI that makes scheduling management instant from the terminal — with a local data layer that enables offline booking search, availability analysis, schedule health monitoring, and team scheduling insights that the web UI makes tedious. gcalcli (3k+ stars for Google Calendar) proves developer demand for calendar CLIs exists.
-
-## API Identity
-
-- **Domain:** Scheduling / Calendar Management
-- **Primary users:** (1) Business owners managing their booking flow, (2) Engineering managers coordinating team availability, (3) Developers building scheduling integrations, (4) Ops/support staff handling booking changes
-- **Core entities:** Bookings, Event Types, Schedules, Calendars, Teams, Organizations, Webhooks, Slots, Attendees, Conferencing
-- **Base URL:** `https://api.cal.com`
-- **Auth:** Bearer token (API key prefixed `cal_live_`), OAuth 2.0 with PKCE
-- **Rate limit:** 120 req/min (API key), upgradeable to 200-800
-- **Versioning:** `cal-api-version` header (date-based, latest: `2026-02-25`)
-- **Pagination:** `take` + `skip` with envelope: `{ status, data, pagination: { totalItems, hasNextPage, ... } }`
-- **v1 deprecated:** Shutting down April 8, 2026
-- **Spec:** OpenAPI 3.0.0 at `docs/api-reference/v2/openapi.json` in repo
-
-### Data Profile
-
-| Dimension | Assessment |
-|-----------|-----------|
-| Write pattern | Mutable (bookings transition: upcoming -> past/cancelled/rescheduled; event types are editable) |
-| Volume | Medium (hundreds to low thousands of bookings per account, tens of event types) |
-| Real-time | Webhooks (outbound only — booking.created, booking.cancelled, etc.), NO WebSocket/SSE for clients |
-| Search need | HIGH — users need to find bookings by attendee, date range, status; find available slots |
-
-## Usage Patterns (Top 5 by Evidence)
-
-| # | Pattern | Evidence Score | What It Needs |
-|---|---------|---------------|---------------|
-| 1 | **Booking management** (create, cancel, reschedule, list upcoming) | 8/10 | CRUD + status filtering + date ranges |
-| 2 | **Calendar sync troubleshooting** (Google/Outlook integration issues) | 7/10 | Doctor command checking calendar connections + busy-time inspection |
-| 3 | **Availability/slot checking** (when am I free? when can someone book me?) | 7/10 | Slot queries with time range + timezone + event type |
-| 4 | **Schedule/availability configuration** (set working hours, block time, OOO) | 6/10 | Schedule CRUD + OOO management |
-| 5 | **Team scheduling management** (who's available, team event types) | 6/10 | Team queries, member schedules, team event types |
-
-**Evidence sources:**
-- Cal.com GitHub issues (#14706: missing endpoints, #25560: workflow API, #3546/#8376: sync issues)
-- n8n/Zapier Cal.com integrations (cross-platform evidence)
-- Official companion CLI structure (46 command categories)
-- Trustpilot reviews citing booking management and sync pain points
-- Cal.com docs featuring availability as core product value prop
-
-## Tool Landscape
-
-### Forge/Platform CLIs
-| Tool | Stars | Language | Type | Commands | Maintained |
-|------|-------|----------|------|----------|------------|
-| @calcom/cli (official, in calcom/companion) | 8 | TypeScript | API Wrapper | 46 categories | Yes (2026) |
-
-### Workflow Overlays
-None developer-native. n8n, Zapier, Make.com provide no-code automation but not CLI workflows.
-
-### Alternative UX (Calendar Domain)
-| Tool | Stars | Language | Relevance |
-|------|-------|----------|-----------|
-| gcalcli | 3k+ | Python | Google Calendar CLI — proves demand exists |
-| calcure | 500+ | Python | TUI calendar/task manager — generic, not API-connected |
-| calcurse | 1.1k+ | C | TUI calendar — local only |
-
-### Data Tools
-None. Zero tools sync Cal.com data locally for offline search/analysis.
-
-### Integration Tools
-- n8n Cal.com trigger (workflow automation platform)
-- Zapier Cal.com integration (no-code)
-- Make.com Cal.com app (visual workflow)
-- cal-pr-agent (5 stars, PR automation for Cal.com dev)
-
-**Key finding:** The Cal.com ecosystem has ZERO non-wrapper tools. No discrawl-equivalent. No offline search, no analytics dashboards, no backup tools. The entire data-tool and workflow-tool lanes are empty. gcalcli's 3k stars for Google Calendar proves developer appetite for scheduling CLIs.
-
-## Workflows
-
-### 1. Daily Agenda
-- **Steps:** GET /v2/bookings?status=upcoming&afterStart=today&beforeEnd=tomorrow → group by time → show attendees
-- **Frequency:** Daily
-- **Pain point:** Must open web UI every morning to see today's meetings
-- **Proposed:** `calcom-pp-cli agenda` — today's bookings with attendee info, times, event types
-
-### 2. Availability Finder
-- **Steps:** GET /v2/slots?username=X&eventSlug=Y&startTime=A&endTime=B → parse windows → format as time blocks
-- **Frequency:** Multiple times daily when coordinating
-- **Pain point:** Need to open Cal.com or share link just to see own availability
-- **Proposed:** `calcom-pp-cli free --days 7` — show free windows for next N days
-
-### 3. Schedule Health Check
-- **Steps:** GET /v2/schedules → GET /v2/bookings?status=upcoming → cross-reference → check for conflicts
-- **Frequency:** Weekly
-- **Pain point:** No way to audit schedule health; sync issues are #1 complaint
-- **Proposed:** `calcom-pp-cli health` — check calendar connections, find conflicts, report utilization
-
-### 4. Booking Analytics
-- **Steps:** Sync all bookings → aggregate by event type → compute show/cancel rates → identify trends
-- **Frequency:** Weekly/Monthly
-- **Pain point:** No built-in analytics; callytics web dashboard is the only option (0 stars)
-- **Proposed:** `calcom-pp-cli stats --days 30` — show rate, cancellation rate, busiest days
-
-### 5. Stale Event Type Cleanup
-- **Steps:** GET /v2/event-types → GET /v2/bookings?eventTypeId=X for each → flag unused
-- **Frequency:** Monthly
-- **Pain point:** Event types accumulate, no easy way to identify unused ones
-- **Proposed:** `calcom-pp-cli stale --days 90` — find event types with zero bookings in N days
-
-## Architecture Decisions
-
-| Decision Area | Need Level | Choice | Rationale |
-|---------------|-----------|--------|-----------|
-| **Persistence** | HIGH | SQLite with domain tables (bookings, event_types, schedules, attendees) | Bookings accumulate; local DB enables offline search, trend analysis, and analytics without API calls |
-| **Real-time** | LOW | REST polling with `afterUpdatedAt` cursor | No WebSocket/SSE for clients; API supports date-range filtering for incremental sync |
-| **Search** | HIGH | FTS5 on booking titles/descriptions, attendee names/emails, event type names | Users need to find bookings by attendee, keyword, date range |
-| **Bulk** | MEDIUM | Paginate with `take=100` + `skip`, incremental sync | 120 req/min rate limit; sync fetches all pages incrementally with cursor |
-| **Cache** | MEDIUM | SQLite IS the cache; read commands query local DB; `--sync` triggers refresh | Avoids rate limits; enables offline analytics |
-
-## Top 5 Features for the World
-
-| # | Feature | Evidence | Impact | Feasibility | Uniqueness | Composability | Data Fit | Maintain | Moat | **Total** |
-|---|---------|----------|--------|-------------|------------|---------------|----------|----------|------|-----------|
-| 1 | **Booking search** (FTS5 across all bookings) | 3 | 3 | 2 | 2 | 2 | 2 | 1 | 1 | **16/16** |
-| 2 | **Agenda** (today's bookings, formatted) | 2 | 3 | 2 | 2 | 2 | 2 | 1 | 0 | **14/16** |
-| 3 | **Schedule health** (gaps, conflicts, utilization) | 2 | 2 | 2 | 2 | 2 | 2 | 1 | 1 | **14/16** |
-| 4 | **Free windows** (show availability) | 2 | 3 | 1 | 2 | 1 | 1 | 1 | 1 | **12/16** |
-| 5 | **Stale finder** (unused event types) | 1 | 2 | 2 | 2 | 2 | 2 | 1 | 0 | **12/16** |
-
-All 5 features score >= 12 — all are must-haves.
-
-## Sources
-
-- [Cal.com GitHub](https://github.com/calcom/cal.com) — 34k+ stars
-- [Cal.com API v2 Docs](https://cal.com/docs/api-reference/v2/introduction) — 181 paths, 285 operations
-- [OpenAPI Spec](https://github.com/calcom/cal.com/blob/main/docs/api-reference/v2/openapi.json) — 1.1MB JSON
-- [Cal.com Companion/CLI](https://github.com/calcom/companion) — 8 stars, 46 command categories
-- [API v1→v2 Migration](https://cal.com/docs/api-reference/v2/v1-v2-differences) — v1 shutdown April 8, 2026
-- [GitHub Issue #14706](https://github.com/calcom/cal.com/issues/14706) — Missing API endpoints
-- [GitHub Issue #25560](https://github.com/calcom/cal.com/issues/25560) — Workflow API request
-- [GitHub Issues #3546, #8376](https://github.com/calcom/cal.com/issues/3546) — Calendar sync problems
-- [n8n Cal.com Integration](https://n8n.io/integrations/cal-trigger/) — Workflow automation
-- [gcalcli](https://github.com/insanum/gcalcli) — 3k+ stars, proves calendar CLI demand
-- [calcure TUI](https://github.com/anufrievroman/calcure) — Terminal calendar alternative
diff --git a/docs/plans/2026-03-27-fix-calcom-cli-audit.md b/docs/plans/2026-03-27-fix-calcom-cli-audit.md
deleted file mode 100644
index 84d162c3..00000000
--- a/docs/plans/2026-03-27-fix-calcom-cli-audit.md
+++ /dev/null
@@ -1,45 +0,0 @@
----
-title: "Non-Obvious Insight Review: Cal.com CLI"
-type: fix
-status: active
-date: 2026-03-27
-phase: "3"
-api: "cal.com"
----
-
-# Non-Obvious Insight Review: Cal.com CLI
-
-## Automated Scorecard Baseline: 55/100 (Grade C)
-
-| Dimension | Score | Gap | Fix |
-|-----------|-------|-----|-----|
-| Output Modes | 10/10 | — | Good |
-| Auth | 8/10 | Missing cal-api-version header | Add header to client.go |
-| Error Handling | 10/10 | — | Good |
-| Terminal UX | 9/10 | No color | Add color later |
-| README | 5/10 | No cookbook, no comparison | Add in Phase 4 P7 |
-| Doctor | 10/10 | — | Good |
-| Agent Native | 8/10 | — | Good |
-| Local Cache | 10/10 | — | Good |
-| Breadth | 8/10 | — | Good (222 commands) |
-| Vision | 8/10 | Generic store | Replace with domain tables |
-| Workflows | 2/10 | No domain workflows | Add 7 workflow commands |
-| Insight | 0/10 | No insight commands | Add stats, stale, conflicts |
-| Sync Correctness | 0/10 | Generic sync | Rewrite with afterUpdatedAt |
-| Dead Code | 0/5 | Dead code present | Audit in Phase 4.6 |
-
-## Critical Issues
-1. Missing `cal-api-version` header — API calls will 400
-2. Generic store — JSON blob, not domain columns
-3. No workflow commands — the product is missing
-4. Wrong binary name — cal-com-user-cli, not calcom-pp-cli
-5. Complex body fields skipped — booking create needs --stdin
-
-## GOAT Plan (Phase 4 priorities)
-- P0: Domain data layer (bookings, attendees, event_types tables + FTS5 + sync)
-- P1: Table stakes (cal-api-version header, CRUD verification)
-- P2: 7 workflow commands (agenda, search, stats, free, unconfirmed, conflicts, stale)
-- P3: Name normalization (calcom-pp-cli) + command name cleanup
-- P4: Scorecard fixes + dead code audit
-- P5: Tests
-- P6-7: Distribution + README polish
diff --git a/docs/plans/dogfood-gauntlet-findings.md b/docs/plans/dogfood-gauntlet-findings.md
deleted file mode 100644
index 68459c80..00000000
--- a/docs/plans/dogfood-gauntlet-findings.md
+++ /dev/null
@@ -1,74 +0,0 @@
-# Dogfood Gauntlet Findings - 2026-03-24
-
-## Scorecard: 10/10 Pass All 7 Gates (was 4/10 before Round 2 fixes)
-
-| # | API | Paths | Result | Notes |
-|---|-----|-------|--------|-------|
-| 1 | Fly.io | 51 | PASS | 7/7 gates. Complex body fields skipped (config, metadata). |
-| 2 | Spotify | 68 | PASS | 7/7 gates. CLI name: spotify-web-sonallux-cli. Array body fields skipped (ids, uris). |
-| 3 | Telegram | 74 | PASS | 7/7 gates. 22 endpoints skipped (resource limit 50). Flat structure. |
-| 4 | Vercel | 85 | PASS | 7/7 gates. Global param `teamId` filtered (93/113 endpoints). oneOf bodies skipped. |
-| 5 | Supabase | 105 | PASS | 7/7 gates. No servers in spec - fallback placeholder URL used. |
-| 6 | Sentry | 126 | PASS | 7/7 gates. Flat "api" resource. Endpoints truncated at limit 50. |
-| 7 | LaunchDarkly | 221 | PASS | 7/7 gates. Flat "api" resource. Endpoints truncated at limit 50. |
-| 8 | Trello | 264 | PASS | 7/7 gates. Global params `key`/`token` filtered (324/324 endpoints). |
-| 9 | Jira | 317 | PASS | 7/7 gates. Flat "rest" resource. Endpoints truncated at limit 50. |
-| 10 | Cloudflare | 1716 | PASS | 7/7 gates. Largest spec tested. Complex body fields skipped. |
-
-## Grade: A (10/10 pass, up from 4/10)
-
-All 6 bugs from Round 1 findings are fixed. The generator handles diverse real-world specs.
-
-## Bugs Fixed (Round 2 - commit 1eda222)
-
-### Bug 1: `$ref`/`$schema` keys leak into Go identifiers - FIXED
-- **Was:** Fly.io, Vercel, Jira failed go vet
-- **Fix:** `toCamel` rewritten to split on all non-letter/non-digit chars. `flagName` rewritten with full sanitization.
-
-### Bug 2: Spec title too long for CLI name - FIXED
-- **Was:** Spotify panic from 60+ char CLI name
-- **Fix:** `cleanSpecName` noise word filtering already handled this (prior commit). Name now derives cleanly.
-
-### Bug 3: Missing `servers` field causes parse failure - FIXED
-- **Was:** Supabase failed with "base_url is required"
-- **Fix:** Parser falls back to placeholder URL when no servers defined (prior commit).
-
-### Bug 4: Enum values with `/` generate invalid Go type names - FIXED
-- **Was:** Trello failed go build
-- **Fix:** `toCamel` now strips `/` and all non-alphanumeric chars. `flagName` does the same.
-
-### Bug 5: `defaultVal` type mismatch - FIXED (new bug found in Round 2)
-- **Was:** Supabase had string "true" used as bool flag default
-- **Fix:** `defaultVal` now coerces defaults by declared param type instead of Go type-switching on the raw value.
-
-### Bug 6: Duplicate body/query param flag names - FIXED (new bug found in Round 2)
-- **Was:** Some specs had body fields that collided with query/path params
-- **Fix:** Parser deduplicates body params against query/path params by kebab-case flag name. Also deduplicates within `mapRequestBody` and `mapTypes` by camelCase name.
-
-### Bug 7: Cloudflare parse failure - FIXED
-- **Was:** "name is required" because spec had no info.title
-- **Fix:** Parser fallback to filename-derived name (prior commit).
-
-## Remaining Observations (Not Bugs)
-
-### Flat Resource Problem
-- Sentry, LaunchDarkly, Jira all produce a single flat resource ("api", "rest") because all paths share a common prefix (`/api/v2/`, `/api/0/`, `/rest/api/3/`).
-- Enhancement opportunity: strip common API path prefixes before grouping into resources.
-
-### Endpoint Truncation
-- Telegram (22 skipped), Sentry, LaunchDarkly, Jira all hit the 50-endpoint-per-resource limit.
-- Consider bumping to 100 for flat-structure APIs, or implementing smarter sub-resource detection.
-
-### Global Param Filtering
-- Vercel's `teamId` and Trello's `key`/`token` correctly auto-detected and filtered as global params.
-- This is working well - the heuristic (present on >80% of endpoints) catches real global params.
-
-## Fix History
-
-| Round | Commit | Pass Rate | Key Fixes |
-|-------|--------|-----------|-----------|
-| Initial | - | 3/10 | Baseline |
-| Round 1 | 3f096bc | 4/10 | Parser type name sanitization, unlocked Jira |
-| Round 1.5 | d8a93e0 | 4/10 | Template sanitization, toCamel/flagName/types hardening |
-| Round 2 | 651d11e | 4/10 | Doctor health endpoint checking |
-| Round 2.5 | 1eda222 | 10/10 | toCamel/flagName full rewrite, defaultVal coercion, body dedup |
diff --git a/docs/plans/overnight-hardening-results.md b/docs/plans/overnight-hardening-results.md
deleted file mode 100644
index 4e07aad9..00000000
--- a/docs/plans/overnight-hardening-results.md
+++ /dev/null
@@ -1,78 +0,0 @@
-# Overnight Hardening Results - 2026-03-25
-
-## Summary
-
-Ran 6 units overnight. 4/6 fully completed, 2 blocked by upstream spec issues.
-
-## Results
-
-### Unit 1: 10-API Gauntlet - PASS (10/10)
-
-All 10 APIs pass 7/7 quality gates with the new templates. Found and fixed 1 regression: `readme.md.tmpl` crashed on APIs with no auth env vars (Fly.io, Telegram) - the `index .Auth.EnvVars 0` call panicked on empty slices. Fixed with conditional guards.
-
-| API | Paths | Result |
-|-----|-------|--------|
-| Fly.io | 51 | PASS (after README fix) |
-| Spotify | 68 | PASS |
-| Telegram | 74 | PASS (after README fix) |
-| Vercel | 85 | PASS |
-| Supabase | 105 | PASS |
-| Sentry | 126 | PASS |
-| LaunchDarkly | 221 | PASS |
-| Trello | 264 | PASS |
-| Jira | 317 | PASS |
-| Cloudflare | 1716 | PASS |
-
-### Unit 2: Tests - 30 new test cases
-
-Created 5 test files with meaningful coverage:
-
-| File | Tests | What's covered |
-|------|-------|----------------|
-| research_test.go | 5 | novelty scoring, dedup, recommendations, round-trip |
-| dogfood_test.go | 4 | tier1 commands, credentials, scoring, round-trip |
-| comparative_test.go | 3 | alt scoring, gap analysis, no-research fallback |
-| textfilter_test.go | 4 | AI slop detection, normal text, formatting |
-| readme_augment_test.go | 3 | markers, no-evidence, append mode |
-
-### Unit 3: PagerDuty CLI - BLOCKED
-
-PagerDuty's OpenAPI spec has a schema reference error: `bad data in "#/components/requestBodies/OrchestrationCacheVariableDataPutResponse"`. Our parser (kin-openapi) is strict about $ref resolution. Needs either a parser fix or a pre-processed spec.
-
-### Unit 4: Plaid CLI - PASS (51 resources, 62 files)
-
-Generated from official OpenAPI spec. All new template features verified: --select flag, error hints with CLI name, generation comments on all Go files, new README format with quickstart. Auth: PLAID_CLIENT_ID + PLAID_SECRET.
-
-### Unit 5: Intercom CLI - BLOCKED, replaced with Pipedrive
-
-Intercom's OpenAPI spec has a schema resolution error: `custom_attributes` component not found. Tried versions 2.10 and 2.11, same issue.
-
-Generated **Pipedrive CLI** instead. 109 files, 7/7 gates pass. CRM from the terminal - deals, persons, organizations, pipelines, activities, leads, and more. Auth via PIPEDRIVE_API_TOKEN.
-
-### Unit 6: Final validation - PASS
-
-All tests pass. 8 commits on main.
-
-## Commits
-
-| Hash | Description |
-|------|-------------|
-| 780c6b5 | feat(templates): --select, error hints, README, Owner variable |
-| 2370b45 | fix(templates): guard readme.md.tmpl against empty Auth.EnvVars |
-| 02d022f | test: 30 new tests for research, dogfood, comparative, textfilter, readme_augment |
-| 66bc4ea | feat(catalog): generate Plaid CLI from official OpenAPI spec |
-| 451b7cb | feat(catalog): generate Pipedrive CLI from official OpenAPI spec |
-
-## Issues Found
-
-1. **README template empty EnvVars crash** - Fixed. APIs with no auth env vars caused `index .Auth.EnvVars 0` to panic.
-2. **PagerDuty spec schema error** - kin-openapi can't resolve `OrchestrationCacheVariableDataPutResponse`. Need to either relax parser strictness or pre-process the spec.
-3. **Intercom spec schema error** - kin-openapi can't resolve `custom_attributes` component. Same class of issue as PagerDuty.
-
-## What's Ready for Morning
-
-- **Plaid CLI** (51 resources) and **Pipedrive CLI** (109 files) are generated and committed
-- Template quality is at ~6.9/10 Steinberger score (up from 4.9)
-- 10-API gauntlet passes 10/10 with new templates
-- 30 new test cases providing first coverage for Phase 1-3 code
-- Two spec parser issues identified for next fix round
← bf14db97 feat(cli): add publish skill to ship CLIs to printing-press-
·
back to Cli Printing Press
·
feat(skills): show existing CLI context and clarify regenera 590a07b1 →