skills/clara/references/workflow-catalog.md

# Clara workflow catalog

Read this catalog before routing professional work. The user describes the
task; Clara selects the workflow. Use semantic judgment, then read the selected
skill completely. Do not select from keywords or use a cross-cutting assurance
skill to imitate a missing operational workflow.

`advisory-case-director` owns durable advisory-case direction: the current
answer, the case-specific analytical spine, the next decision-relevant work,
the integration of returned evidence and partner judgement, and the timing of
working deliverable updates. The main `clara` skill routes work and provides
shared case-workspace mechanics. Use the other specialist workflows below when
their bounded purpose matches the requested contribution.

## Professional workflows

- `advisory-brief-planner`: turn a natural advisory assignment into the
  reviewable `advisory_contract.json` assignment and generation-handoff
  contract, then route execution to the existing Clara workflow that retains
  procedural authority.
- `advisory-case-director`: direct a durable advisory case after assignment
  framing; state the answer first, maintain the partner-readable analytical
  spine and cumulative evidence lineage, choose the next research, interview,
  data, or deliverable branch, integrate what returns, and revise the position
  when evidence or partner judgement warrants it.
- `advisory-deliverable-validator`: validate a completed advisory memo, report,
  analysis, presentation, or other supported professional document by walking
  generation-time evidence and claim dependencies when available, or by
  explicitly labelled matched support for an external document; assess it
  against `advisory_contract.json` while composing with rather than duplicating
  the applicable format-specific Clara checks.
- `attribute-reporting`: map retailer products to a central taxonomy, compare a
  retailer-defined cohort with the remaining assortment, and create a checked
  private Retailer Signals report.
- `brand-fit`: compare checked retailer signals with the brand's current
  retailer presence and brand-owned catalogue in the stored snapshot.
- `business-planning`: prepare one business plan for a startup, new venture or
  established company. Assess customers, market, operations, economics, cash,
  options, recommendation and next actions using one case, financial model and
  report. Vera and Clara expose this exact same function. The business question
  determines scope; the entry product never changes the angle or required work.
- `deck-correction`: turn spoken or written feedback into verified changes to
  an existing PPTX or Clara HTML deck while preserving untouched content.
- `html-deck`: build or revise a source-faithful, standalone HTML presentation
  with fixed-stage rendering, motion, navigation, and browser QA.
- `hosted-interview`: prepare and operate an adaptive hosted external interview,
  including brief selection, participant link, retrieval, and quality review.
- `research-video`: turn an approved ordered set of research scene images into
  a source-faithful 16:9 narrated MP4 with restrained motion, synchronized
  narration generated through authenticated Mparanza access in English, Italian, French, German, or Spanish, a
  visible localized AI-voice disclosure, captions, a narration script, and
  mechanical media QA, without a user API key. Mparanza sends only the approved
  narration and binding hashes to OpenAI and returns downloadable scene-level
  audio artifacts. Accepted Vera outputs may be
  handed off without replacing Vera's professional source and claim review.
- `reporting-engine`: review dataset semantics, select meaningful business
  analysis, and prepare deterministic calculations, charts, or reporting
  evidence from CSV, XLSX, or Parquet inputs.
- `transcribe`: capture, transcribe, attribute, import, and review advisor voice
  notes, debriefs, meetings, or calls.

## Cross-cutting assurance

- `claim-basis-map`: create or audit a readable sidecar mapping presentation
  claims to sources, calculations, prior claims, reasoning, assumptions, or an
  explicit ungrounded status.

Use this assurance workflow when its own trigger conditions apply. It does not
replace a missing presentation, research, analysis, or advisory workflow.

## Developer governance

- `privacy-surface-review`: review and register model-context and external-data
  boundaries when a Clara workflow or shared service changes. This is not a
  customer-case workflow.

<!-- CLARA_OPENAI_ONBOARDING_BEGIN -->
## Learning

- `learn-with-clara`: teach current workflows by native voice with a parallel visible working chat, first-use interview, 3–4 tailored lessons and reusable local examples.
Only Clara's own installed operational skills are eligible. Explain requests
for another plugin's skills and offer relevant Clara workflows without teaching
or silently substituting the other plugin.
<!-- CLARA_OPENAI_ONBOARDING_END -->