Read and apply docs/COLLABORATIVE-DESIGN-PRINCIPLE.md, standards/evidence-lifecycle.md and standards/notes-adr-sync.md for scoped authority, exact evidence and decision ownership. Existing named authority continues; analysis is read-only and report-only excludes input/index/state writes.

User Guide

• When to use: Generate a project progress dashboard from workflow-catalog.yaml. Reads current stage, required steps, artifact evidence, validation gaps, writes production/project-roadmap.md after approval, and mirrors to memory_bank/t2_execution/current_roadmap.md only when the Memory Bank root exists and that path and effect are covered.
• Inputs: Command arguments: /cdd-status [optional: --dry-run | --write]; project artifacts referenced below; user decisions and approvals before writes.
• Outputs: Primary artifacts, reports, or conversation guidance described below; write files only after user approval.
• Memory-bank reads: existing laws/current-state/workflow-contract/adapter-state.
• Memory-bank writes: only with initialized Memory Bank and the separately covered memory_bank/t2_execution/current_roadmap.md mirror; no law/current-state/workflow-contract/adapter-state writes.
• Next steps: Follow the workflow hand-off or next-step guidance below; recommendations do not auto-run and require explicit user command/approval.

CDD Status Dashboard

Generate a concise progress dashboard and a durable roadmap at production/project-roadmap.md. For the expected shape, see docs/examples/project-roadmap.example.md.

When memory_bank/ exists and its exact mirror effect is covered, maintain memory_bank/t2_execution/current_roadmap.md. This mirror is for project memory and does not replace production/project-roadmap.md.

memory_bank/t2_execution/adapter_state.yaml is recorded by /constitute-check. /cdd-status reads and displays that record but never recomputes freshness and never writes or owns the adapter-state file.

This skill bridges /help and /project-stage-detect:

• /help gives one next required step.
• /project-stage-detect performs a full audit.
• /cdd-status creates a saved progress dashboard from the catalog.

Collaboration Rule

Draft the roadmap and disclose actual paths/effects first. Reuse existing named roadmap authority across retries. Missing scope asks: "May I write production/project-roadmap.md and the named memory_bank/t2_execution/current_roadmap.md mirror?" The named mirror may be maintained or created when the Memory Bank root exists and that path and effect are covered. The mirror is a distinct effect, never a hidden gift from report-only approval. An explicit --write flag or write instruction covers only its documented requested effects; disclose them and ask only for any material new scope. --dry-run always writes nothing. Without Memory Bank, write only the covered roadmap and report mirror skipped; no initialization is required or performed. Optional advice when governance setup is requested: "Run /constitute to establish the memory_bank governance control plane". This is a suggestion only: do not invoke it, initialize Memory Bank, change default QA handling, or make independent reporting/qualification depend on optional setup. All other law/contract/adapter/session/sprint/stage/index effects remain excluded.

1. Read Authoritative Inputs

Read these files when present:

• workflow/workflow-catalog.yaml
• production/stage.txt
• production/sprint-status.yaml
• production/session-state/active.md
• memory_bank/t0_core/basic_law_index.md
• memory_bank/t0_core/current_state.md
• memory_bank/t2_execution/workflow_contract.md
• memory_bank/t2_execution/adapter_state.yaml
• design/cdd/game-concept.md
• design/cdd/product-concept.md
• design/ux/surface-profile.md

Detect domain from substantive concept bodies and configured values under standards/technical-preferences.md, including valid populated legacy Product aliases. Report Game, Product or Unknown for the current scope. Conflicting concepts or mixed scope need a concrete Game or Product selection before dependent catalog filtering/routing; do not auto-select Both or qualify completion from this choice. Missing concepts never default Game; actual legacy Product remains Product. Independent neutral reporting continues.

Report four facts independently:

1. Declared phase: actual production/stage.txt content or owner statement, including conflicts.
2. Observed bodies/locators/status projections and omitted inputs.
3. Candidate phase: advisory ongoing-work indicators under /project-stage-detect.
4. Qualified state: only exact governing checks/decisions/reviews/transition or completion authority actually verified; otherwise unverified.

Detect recorded adapter freshness when the state file exists:

• Read status, checked_commit, checked_at, manifest_digest, and source_digest.
• Display fresh, stale, or uninitialized as a recorded value, not a live check. If the file is absent, display not initialized.
• For stale, uninitialized, or missing state, add a risk recommending /constitute-check.
• Do not change the catalog-derived blocker or reorder the next three commands because of adapter state.

2. Parse The Workflow Catalog

Use workflow/workflow-catalog.yaml as the only required-step source of truth.

For each phase:

• Keep steps with no applies_to.
• Keep applies_to: [game] only for Game projects.
• Keep applies_to: [product] only for Product projects.
• For Unknown projects, label domain-specific steps as applicability pending. Hold dependent single-domain routing/completion claims until a concrete Game or Product scope resolves; independent neutral reporting continues.

For each step, extract:

• id
• name
• command
• required
• required_when
• artifact.glob
• artifact.min_count
• artifact.note
• description

3. Check Completion

Classify every required step:

Status	Meaning
COMPLETE	Actual governing required checks/decisions/reviews/completion authority verified against exact evidence
PARTIAL	Artifacts observed, but required content/evidence or qualification incomplete/unverified
MISSING	No artifact or manual evidence was found
N/A	required_when is false and a rationale exists, usually in design/ux/surface-profile.md
MANUAL	The step has no machine-checkable artifact; report what must be verified

Rules:

• Globs/counts/patterns are catalog structural observations. Read substantive bodies and actual required dependency closure before qualified COMPLETE; missing/failed/unexecuted required checks retain their actual state.
• Resolve explicitly referenced direct/indirect inputs to their actual paths and read their bodies, including ignored or untracked attachments. A discovery glob or ignore-aware search omission does not establish absence. Check each claimed missing path directly and retain its individual result; a silent test in an aggregate command is not a path-specific absence observation. Confirmed absence, unread inputs and inaccessible inputs remain distinct. Unread or inaccessible required inputs keep the affected check Pending/unverified.
• If a step has required_when, evaluate it from the domain and design/ux/surface-profile.md when available.
• If applicability is ambiguous, mark the step MANUAL, not COMPLETE.
• Catalog governs workflow step ordering/applicability; owning CDD/Story/decisions govern actual required evidence. Do not invent requirements or erase real ones.

4. Determine The Blocker And Next Commands

Find the first incomplete required step in the current phase.

If all current-phase required steps are complete:

• The current blocker is the phase gate, e.g. /gate-check technical-setup.
• The next commands are the first required commands in the next phase.

If a required step is incomplete:

• The current blocker is that step.
• The first next command is its command.
• The next two commands are the following required commands in phase order.

5. Build The Roadmap

Write a roadmap with this structure:

# Project Roadmap

> Generated by `/cdd-status` from `workflow/workflow-catalog.yaml`.
> Last updated: [date]

## Snapshot

- Domain: [Game/Product/Unknown]
- Declared phase: [source/value or absent]
- Observed work / candidate phase: [body evidence and advisory candidate]
- Qualified state: [verified scope/evidence or unverified]
- Catalog structural observations: [observed / actual applicable denominator]
- Qualified required progress: [verified complete / actual applicable denominator]
- Current blocker: [step or gate]
- Adapter freshness (recorded): [fresh/stale/uninitialized/not initialized]
- Adapter state checked: [checked_at at checked_commit, or never]

## Next Commands

1. `/command`
2. `/command`
3. `/command`

## Phase Progress

| Phase | Applicable Required | Observed | Qualified Complete | Missing/Pending |
| ----- | ------------------- | -------- | ------------------ | --------------- |
| ... |

## Current Phase Checklist

| Step | Required | Evidence | Status |
| ---- | -------- | -------- | ------ |
| ... |

## Product Surface Decisions

| Artifact | Status | Source |
| -------- | ------ | ------ |
| `design/ux/interaction-patterns.md` | REQUIRED / N/A / MISSING | `design/ux/surface-profile.md` or catalog |
| `design/design-system.md` | REQUIRED / N/A / OPTIONAL | surface profile / scope |
| `design/brand/style-guide.md` | OPTIONAL / REQUIRED BY SCOPE / N/A | surface profile / scope |

## After This Phase You Should Have

- [artifact glob or manual evidence generated from the catalog]

## Risks

- [missing critical evidence, stale sprint data, unresolved gate failure, or N/A without surface profile]
- [stale, uninitialized, or missing adapter state; recommend /constitute-check]

## Notes

- Catalog is authoritative.
- A risk acceptance/exception is separate; it cannot turn required FAIL/NotRun/
  Blocked/Pending/Unknown into PASS or qualified completion/transition.

Keep the console response under 60 lines. The saved roadmap may be longer.

When writing the T2 mirror, use the same roadmap content with this extra header:

> Governance memory mirror generated by `/cdd-status` from
> `workflow/workflow-catalog.yaml`.
> Customer/team-facing roadmap: `production/project-roadmap.md`.

6. Product Surface Profile Handling

For Product projects, read design/ux/surface-profile.md when present.

Use it to decide whether these are applicable:

• design/ux/interaction-patterns.md
• design/design-system.md
• design/brand/style-guide.md

Populate the roadmap's Product Surface Decisions table for Product projects:

• interaction-patterns.md: REQUIRED when the product has any API, CLI, SDK/library, UI, admin, operator, docs-driven consumer journey, or other user/integrator-facing surface; N/A only with surface-profile rationale; MISSING when required and absent.
• design-system.md: REQUIRED for UI-heavy or multi-surface UI products, N/A for API-only, CLI-only, SDK/library, or internal headless products with rationale, otherwise OPTIONAL.
• style-guide.md: REQUIRED BY SCOPE when public brand, docs imagery, screenshots, diagrams, marketing/release visuals, or visual tone are in scope; otherwise OPTIONAL or N/A when explicitly ruled out.

If a project marks one of these N/A without design/ux/surface-profile.md, flag a risk:

Product surface requirement marked N/A without design/ux/surface-profile.md.

Recommend creating it from templates/surface-profile.md.

7. Output

Always report:

• Domain
• Declared/observed/candidate/qualified phase facts
• Structural observations and qualified progress with actual denominator
• Current blocker
• Recorded adapter freshness and its checked commit/time, or not initialized
• Next 3 commands
• Product surface decision table when domain is Product
• Whether production/project-roadmap.md was written or only drafted
• Whether memory_bank/t2_execution/current_roadmap.md was written, skipped because memory_bank/ is missing, or skipped because this was --dry-run or outside covered mirror authority

Do not mark any manual or structural step qualified complete without its actual required exact evidence/authority. A recorded done status needs owning closure recheck. Do not describe recorded adapter state as live evidence and do not let it override the workflow catalog's phase ordering.

Recorded source_digest covers only manifest-declared classes; linked neutral owners need independent exact identities. Displayed freshness is historical context, not live runtime/skill/spec/category qualification. Runtime availability follows actual metadata and adapters/README.md, not source files or adapter-state values.