Prompt 00 — Repository discovery and baseline

Attach or reference

• AGENTS.md if already present
• The supplied screenshot pack and descriptions
• The adaptive framework pack
• Existing README, architecture docs, package manifests, deployment files, and test configuration

Prompt

You are conducting the first enterprise-uplift pass for Capablio. Do not implement, refactor, rename, install dependencies, change schemas, or modify runtime behavior in this step.

Inspect the repository thoroughly and build a factual baseline. Use non-mutating commands unless running an existing test/build command is safe. Do not infer a framework or service merely from filenames; confirm it from manifests and code.

1. Map the repository

Identify and document:

• Monorepo/workspace structure
• Frontend applications and frameworks
• Backend services, jobs, workers, and functions
• Package manager and build system
• Shared UI/component packages
• Routing model and public/authenticated route boundaries
• Authentication/session providers
• Tenant-resolution model
• Authorization model and where checks occur
• Database, migrations, ORM/query layer, and tenant-isolation mechanism
• Object storage, queues, email, analytics, logging, monitoring, and error tracking
• Encryption/key-management implementation
• Existing feature flags and entitlement checks
• Existing docs, ADRs, OpenAPI/schema files, generated clients, and release notes
• Test suites and CI/CD pipelines
• Deployment environments and configuration pattern

2. Map current user experiences

Create an inventory of all existing routes and classify each as:

• Public marketing
• Documentation/help
• Authentication/onboarding
• Individual workspace
• Manager workspace
• HR workspace
• Tenant administration
• Global platform operations
• External-rater flow
• Development-only or diagnostics

For each route record:

• Route
• Page/component entry point
• Persona/role
• Tenant/scope assumptions
• Server-side permission check
• Data classes accessed
• Empty/loading/error behavior
• Responsive status
• Test coverage
• Screenshot correspondence where available

3. Compare the code to the supplied screens and framework

Record which screenshot capabilities are:

• Fully implemented
• Fixture/mock only
• Partially implemented
• Missing
• Implemented differently from the screenshot description

Map the adaptive framework concepts already present in code:

• Construct, question family, variant, pack, bundle
• Cadence and schedule
• Exposure cooldown
• Exposure ledger
• Anchor/rotation/PDP composition
• Coverage cycle and feasibility
• Tenant overlays and versions
• 180/360 relationships and thresholds
• Imported Thomas evidence
• PDP recurrence overrides
• Entitlements and active-subject seats

4. Establish the quality baseline

Run existing checks where safe and report exact commands and results:

• Type checking
• Linting
• Unit/integration tests
• End-to-end tests
• Build
• Documentation build
• Accessibility tests if present
• Security/static analysis if present

Do not fix failures. Separate pre-existing failures from inability to run due to missing local services or secrets.

5. Create documentation outputs

Create or update only these documentation files:

wiki/internal/architecture/repository-baseline.md
wiki/internal/architecture/current-system-map.md
wiki/internal/architecture/baseline-test-results.md
wiki/internal/product/current-capability-gap.md
wiki/shared/route-map.md
wiki/shared/decision-log.md

If the dual wiki does not exist yet, create only the minimal folders and files needed for these outputs; prompt 01 will formalize the structure.

Required content

repository-baseline.md must contain:

• Confirmed stack table
• Repository tree summary
• Commands for local development and tests
• Environment/service dependencies
• Unknowns and blockers

current-system-map.md must contain:

• Mermaid context/container diagram
• Request/session/tenant flow
• Primary data stores and jobs
• Trust boundaries
• Current encryption path

current-capability-gap.md must contain:

• Screenshot-by-screenshot implementation status
• Public marketing/docs gap
• Role-workspace gap
• Adaptive framework gap
• Enterprise administration/operations gap
• Prioritized risk list: security/privacy, data integrity, usability, commercial credibility, maintainability

Stop conditions

Stop and report rather than guess if:

• The uploaded pack is not present in the workspace and no equivalent source exists.
• There are multiple application roots with no clear production entry point.
• Tenant resolution or authentication cannot be determined.
• Baseline tests require destructive setup.
• Credentials or production access would be needed.

Completion response

Use the response format required by AGENTS.md. Include a concise list of the ten highest-risk findings and the exact documentation files created. Do not propose implementation estimates yet.