Date: 2026-09-26. Scope: authored Crew and Automation Playbook library in this checkout. This is a progress audit, not a claim of customer or production readiness.

Requirement from the launch catalog	Current proof	State
Browse coverage for Finance, Marketing/Website Growth, Sales, Customer Success, Customer Support, Product, Operations, Engineering, QA, Security, GTM, and Shopify	docs/design/crew_template_catalog.md; generated playbooks/crew-agents/*/catalog.json	All named categories have locally installable Crew definitions.
Reusable single-Crew setup	71 generated Crew records with selected skill and pending chat checklist	Available locally; no customer setup is implied.
Multi-Crew proposal and typed handoffs	40 locally installable multi-Crew Playbooks inside 55 total packages; 45 contract suites pass python3 playbooks/scripts/validate_playbooks.py. The Playbook detail action sends Builder each authored handoff and asks it to set the producer Crew's exact JSON context_output, apply a structural validation_schema, then run the bundled business validator as a blocking step before its consumer. Crew runtime now checks a configured schema before completion and retains a failed output log; a two-Crew unit test rejects malformed and missing-field artifacts before the consumer call.	Authored and mechanically checked. Builder must still insert and demonstrate the route-specific business validator in an actual Workflow.
Trial account to reviewed sales assist	Trial Account to Reviewed Sales Assist v0.1.0; Product Adoption Analyst v3 → Sales Follow-up Coordinator v3, exact trial/subscription, product-event and CRM joins, coverage-aware use observation, current contact gates, unsent draft or no-contact/needs-review result, rejected inactivity/blocked-contact/false-send cases and ten setup checks	Fictional typed route; real trial/event coverage, CRM permission and seller approval remain unproved.
Generic order exception to reviewed customer update	Order Exception to Reviewed Update v0.1.1; Order Operations Coordinator → Support Reply Drafter, exact business/order/customer/case artifact and source-record joins, label versus carrier acceptance distinction, unsent case-linked draft, rejected false delivery/refund/resolution and ten pending setup checks	Fictional typed route; real order, carrier, case and contact-policy sources and Support owner review remain unproved.
Access exception to verified fix	Access Exception to Verified Fix v0.1.0; Access Review Analyst → Security Remediation Coordinator, exact approved policy cell, direct server observation, reviewed change, affected deployment, independent same-cell retest, rejected merge-only closure and ten setup checks	Fictional typed route; real written scope, authorized test identity, current policy and deployed retest remain unproved.
Signed Sales to CS handoff	Signed Deal to Onboarding Handoff v0.1.0; Deal Follow-through Coordinator → Customer Onboarding Coordinator, executed-agreement and entitlement joins, receiving-owner acceptance, blocked and rejected cases, ten setup checks. New Customer to First Value v0.2.0 validates the accepted artifact's exact account, scope, goal, rule and target before its onboarding register proceeds.	Fictional typed route; real agreement, current entitlement, receiving-owner decision and downstream onboarding remain unproved.
Renewal readiness to owner decision	Renewal Risk to Owned Decision v0.1.0; Customer Health Coordinator → Renewal Coordinator, exact account and health join, executed contract/subscription re-read, timezone-aware notice calculation, pending/unknown/missing-terms cases, rejected false renewal/action and ten setup checks	Fictional typed route; a real agreement, billing account, prior notice and renewal-owner decision remain unproved.
Released feature to adoption decision	Released Feature to Adoption Decision v0.1.0; reused Product Adoption Analyst → Product Feedback Coordinator, release/flag and denominator-bound baseline, comparable and incomplete observations, current issue review, rejected false actions, ten setup checks and installer test	Fictional typed route; a real deployed feature, product analytics join, issue read and product-owner review remain unproved.
Vendor evaluation to purchase decision	Vendor Evaluation to Purchase Decision v0.1.0; Vendor Researcher → Spend & Payables Coordinator, exact quote/plan and must-have evidence, term-cost arithmetic, current duplicate/budget/security/privacy gates, pending/approved/blocked examples, rejected fake purchase and ten setup checks	Fictional typed route; a real quote, customer vendor register, budget and owner decision remain unproved.
Engineering operations metric to owned improvement	Engineering Operations Intelligence v0.2.0 promotes its Workflow guide to Engineering Operations Analyst → Engineering Delivery Coordinator, comparable/baseline/unknown metric evidence, current issue review, rejected false cause/ranking/action, ten setup checks and installer test	Fictional typed route; a real engineering source join, owner decision and later independent observation remain unproved.
Website Growth from proposal to measured result	Website Growth Loop v0.5.0, approved page, ship, distribution and traffic examples	Fictional complete route; source truth and a real site run remain unproved.
FinOps cost to billed savings	Cost Anomaly to Verified Savings v0.6.0; Cloud Cost Analyst → Engineering Delivery Coordinator → Finance Analyst, with pending change and verified billing fixtures, a rejected false savings claim, ten setup checks and an installer copy test	Fictional typed route; a real billing source, change approval, deployment and later bill remain unproved.
Post-incident learning to verified improvement	Post-Incident Review and Actions v0.5.0; new Post-Incident Reviewer → Improvement Follow-Through Coordinator, approved/draft review, pending/verified actions, rejected unsupported cause and closed-issue completion, ten setup checks and installer test	Fictional typed route; a real incident, review decision, issue receipt and independent control check remain unproved.
Signup-to-paid funnel	Funnel and Conversion Intelligence v0.2.0; new Funnel Analyst → Growth Experiment Planner, reconciled or baseline-first stage counts, pending plan and rejected false winner, ten setup checks and installer test	Fictional typed route; real event/billing join, owner decision and later experiment outcome remain unproved.
SaaS activation and retention	Activation and Retention Intelligence v0.2.0; new Lifecycle Analyst → Growth Experiment Planner, comparable, baseline-first and pending-maturity cohorts, pending plan and rejected false winner, ten setup checks and installer test	Fictional typed route; real event/billing join, owner decision and later experiment outcome remain unproved.
Approved experiment to outcome	Growth Experimentation and Follow-Through v0.3.0; new Experiment Run Coordinator → Growth Outcome Analyst, pending approval and window, exact provider launch, inconclusive and measured readouts, rejected false launch/winner, ten setup checks and installer test	Fictional typed route; real plan approval, provider and outcome sources, owner decision and later action remain unproved.
Sampled AI-answer visibility	AI Visibility Intelligence v0.2.0; AI Visibility Analyst → Search Opportunity Mapper, two-answer snapshot, pending page opportunity, rejected inflated citation/publication claim, ten setup checks and installer copy test	Fictional typed route; real answer access, page facts, owner review and repeat sampling remain unproved.
Focused SEO review	SEO Intelligence v0.2.0, SEO Analyst → Search Opportunity Mapper, dated issue-to-question artifacts and rejected publication claim; mocked installer test copies its ten-check setup and validator	Fictional typed route; real Builder run, page/source and owner review remain unproved.
SaaS finance collections, payables and refunds	Finance Operations Review, Subscription Receivable to Verified Outcome, Invoice Intake to Reviewed Payable, Refund Request to Reconciled Outcome	Exact-object fictional contracts; actual provider, bank and ledger integration remains customer-specific.
Payment dispute evidence to observed finance state	Dispute to Reconciled Outcome v0.1.0; Billing Operations Coordinator with Dispute Review → Revenue & Close Analyst, exact account/dispute/payment/case revision, timezone-aware deadline, reason-relevant evidence, approved submission receipt, provider outcome, and separate principal/fee ledger gate. Pending and rejected false submission/win examples pass the contract suite.	Fictional typed route; a real processor case, authorized evidence, owner decision and ledger observation remain unproved.
Shopify operator journeys	Seven Shopify Crews and seven Automation Playbooks, with source and rejected examples	Fictional contracts; no live merchant run is proved.
Production availability	Current branch is not the planned final Dominion deployment	Pending by user direction: deploy at the end.

The launch catalog's quality rule asks each Crew for an inspectable input/output example, a failed example, a hard source probe, owner review, repeat rule and a customer-like exercise. A passing package validator proves shape and links, not this full content standard. The seven Engineering, five Customer Success, three inbound Sales, and three core Finance Crews now include worked and rejected outputs plus role-specific source probes in their installed skills. Shopify skills have illustrative output and rejection sections. The three QA Crew templates now have v2 source probes and acceptance checks for pinned journey attempts, original failed and retry traces, and exact release-policy suite joins. The four Customer Support Crews now have v2 source probes and negative setup conditions for SLA/duplicate triage, current cited replies, receiving-owner escalation acceptance and feedback denominator/privacy checks. The three Security Crews now have v2 probes for deployed finding applicability, direct server permission evidence, and exact approved-change/deployment/retest joins. Six Marketing Crews now have v2 source and acceptance checks for comparable competitor captures, campaign/CRM joins, paid funnel stages, frozen experiment plans, provider-confirmed launch and mature outcome readouts. The two GTM Crews now require buyer/offer evidence and provider/lead identity checks. The six Operations Crews now have v2 source probes and negative setup conditions for goal coverage, meeting action acceptance, milestone completion, carrier state, vendor must-haves and document duplicate checks. Product Feedback Coordinator v3 now has distinct source and acceptance probes for feedback themes and feature adoption, with setup scoped to the selected route. These are authored checks, not evidence of a customer run. The next highest-value gaps are:

1. Remaining category-by-category content review: The newly deepened skills meet the example/probe pattern, while the other installed roles and older single-workflow Playbooks still need an individual review against the same rule before the library is called complete.
2. Customer case verification: Authored examples now cover the core Finance, inbound Sales, Customer Success, and Engineering roles, but no actual account, source access or owner review has been exercised. The same applies to the other template categories until pilot setup evidence is recorded.

These are prioritized content gaps, not proof that the remaining skills are complete. Review each category's actual installed skill and setup guide before promotion. Exercise at least one authorized customer-like case per role during pilot setup. Keep external sends, payments, infrastructure changes and schedules behind the customer's reviewed action route.

The package validator now requires every multi-Crew Playbook to carry five to ten tracked setup checks, a rejected example artifact and an executable handoff suite. This is a minimum authoring gate; it does not judge source truth, whether a negative fixture is representative, or whether a customer has completed setup.

The other 15 of 55 installable Playbook packages have no agent_slots, SETUP.json, or executable handoff suite. They remain Workflow guides in the same catalog; they are not yet equivalent to the 40 chat-led multi-Crew proposals. The Playbook picker now labels the two kinds separately and states that Workflow guides have no predefined Crew team or tracked setup checklist; an installer test confirms this for Basic Browser Setup. This count excludes the example package under playbooks/templates/.

Area	Packages	Launch disposition and reason
Browser QA (8)	Basic Browser Setup; Authentication and Session Validation; Role and Permission Validation; Critical Journey Validation; Flaky-Test Detection and Stabilization; Browser Test Self-Healing; Scheduled Regression and Synthetic Monitoring; Release and PR Quality Gate.	Keep as specialized Workflow guides for setup, test methods and recurrence. Browser Journey QA Analyst, Flaky Test Investigator and Release Quality Assistant are the installable Crews; Release Candidate to Reviewed Gate is the typed team route. A guide does not imply that its test runner, schedule or approval path is configured.
Growth Analytics foundation (1)	Growth Data Foundation.	Keep as a shared data/identity setup guide. It is useful beneath multiple Crew jobs but does not itself require a second Crew.
Reliability Operations methods (3)	CI and Deployment Failure Triage; Incident Investigation and Coordination; Governed Remediation and Recovery.	Keep as focused investigation/remediation Workflow guides. Incident Investigator and Engineering Delivery Coordinator can use their methods; Incident to Verified Recovery is the typed team route. Never imply a remediation action follows from installing a guide.
Performance Engineering (2)	Browser Performance Validation; API Performance Validation.	Keep as measurement methods for Performance Investigator. A second Crew is useful only when an actual owner or access boundary is identified.
Security Engineering (1)	Application Security Assessment and Remediation.	Keep as an authorized assessment method guide. Security Findings Analyst and Security Remediation Coordinator already form the typed Finding to Verified Remediation route; the broad assessment package does not itself configure those Crews.

These dispositions describe the current installable product, not a claim that the guides are operationally configured. Continue category-by-category content review and customer-like pilot cases before launch promotion.

After the remaining content and legacy package decisions, run generated catalog checks, Playbook validation, targeted Crew tests and a production frontend build. Then inspect a real Builder installation: Playbook proposal, Crew reuse/creation, actual file paths, blocking validator behavior, setup evidence, owner review and manual first run. Deployment and admin-only Dominion testing come after the library and installation checks.