Draft governance-conformant L1/L2/L3 capabilities for a target industry. Three modes — Industry (user names ONLY an industry; skill proposes the full L1 set), New L1 (user names one new L1), Extend (user names an existing L1 to add children to). The user always speaks in NAMES; the skill resolves names to IDs internally and drives scripts/cli/add.ts.
Synthesise a governance-conformant business-process tree (BP1 + L2/L3) from one Cross-Industry Value Stream and the Business Capabilities its stages exercise. Two modes — Synthesize-from-VS (primary) when the user names a value stream; Extend-VS (maintenance) when the user names a VS already represented as a BP1 plus a new stage. The user speaks in NAMES (VS name, stage name); the skill resolves to ids internally and drives the bp:bootstrap-bp1 + bp:add CLIs.
Draft a governance-conformant value stream (VS) with stages linked to L1 capabilities and (optionally) business processes. User names the trigger and outcome (e.g. "Quote-to-Cash") plus industry applicability; skill drafts stages, resolves capability and process names to ids internally, and drives scripts/cli/vs_add.ts and vs_add_stage.ts. Supersedes /map-value-streams.
Translate the catalogue into a target language by generating sidecar files under catalogue/i18n/<locale>/. Handles three artefact kinds — capability (L1-*.yaml), business-process (processes/BP1-*.yaml), value-stream (_value-streams.yaml). Two scope modes — Whole-Language (user names ONLY a language; skill produces or refreshes a sidecar for every L1, BP1, and the VS file) and Single-Source (user names one source artefact + a language). User speaks in language NAMES and source NAMES; skill resolves to BCP-47 tags and BC/BP/VS ids internally.
Autonomously propose stages in catalogue/_value-streams.yaml linking L1 capabilities (or all L1s of an industry) to end-to-end value streams. Stages map at L1 only; the skill decides which streams apply, flags coverage gaps when no stream fits, and presents ONE batched proposal — no per-stage interrogation. The user always speaks in NAMES; the skill resolves names to IDs internally.