feat(lifecycle): CAM double-rest & monocarpic-kindel phase-sequence fine-typing (REQ-003 D9–D12)

[nolte]

Context

Follow-up carved out of the now-complete perennial-lifecycle Epic #565 (WP-1..8 merged). Phase 1 (#601) wired the cyclic perennation flow and bound the ~90 path-A perennials onto cyclic templates, but deliberately left the CAM-succulent and monocarpic cohorts on the generic evergreen_foliage_perennial / indoor_default sequences rather than their biologically precise flow. Phase 3 (#608) cleaned the dormancy-gap master data but did not add the specialised CAM/monocarpic sequence semantics. REQ-003 already documents the needed templates in its changelog (D9 CAM double-rest, D10 monocarpic kindel/clonal continuation, D11 photoperiodic ornamental induction, D12 palm/fern/geophyte).

Goal

Give the succulent/bromeliad/photoperiodic-ornamental/geophyte cohorts their biologically precise phase-sequence models instead of the generic fallbacks, seeding the pre-formulated templates from the #576 audit.

Scope

• Seed the missing phase-sequence patterns pre-formulated in spec/analysis/seed-phase-sequence-audit.md (schema-valid drafts already exist there): CAM double-rest, monocarpic-kindel/clonal continuation, photoperiodic-ornamental induction, palm/fern/fine-geophyte.
• Re-assign the affected indoor species (CAM succulents, monocarpic bromeliads/agaves, poinsettia/kalanchoe photoperiodics, palms, ferns, geophytes) from indoor_default/evergreen_foliage_perennial onto the precise sequence.
• Idempotent/additive rebind migration for existing installs (claim the next free version at merge; v0024 is the last on develop).
• Consider the natural next step noted in Audit seed species ↔ phase sequences: verify appropriate Phasenabläufe, list unmatched species, pre-formulate missing sequences #576's out-of-scope: an attribute-driven resolver (cycle_type + flowering_strategy + family → template) to replace the static indoor_default blanket — at minimum evaluate/flag it.

Acceptance criteria

• The pre-formulated CAM/monocarpic/photoperiodic/palm-fern-geophyte sequences are seeded (schema-valid, DE+EN naming) and the relevant species reference them.
• No affected species relies on indoor_default purely for lack of a better pattern (verify against the Audit seed species ↔ phase sequences: verify appropriate Phasenabläufe, list unmatched species, pre-formulate missing sequences #576 audit gap list).
• Monocarpic species terminate + continue clonally (D10) rather than cycle-restart; CAM species get their double-rest; where D10 clonal continuation (feat(lifecycle): D10 clonal continuation for monocarpic mothers #390) applies it is honoured.
• Additive rebind migration; seed-schema validation + tests green; REQ-003 changelog bump.

References

• Audit spec/analysis/seed-phase-sequence-audit.md (Audit seed species ↔ phase sequences: verify appropriate Phasenabläufe, list unmatched species, pre-formulate missing sequences #576 — pre-formulated draft sequences + gap list)
• spec/req/REQ-003_Phasensteuerung.md (D9–D12), spec/analysis/lifecycle-flow-completeness-audit.md
• Epic Epic: implement perennial outdoor lifecycle model (wire cyclic restart + per-instance cultivation cycle) #565 (Phase 1 feat(lifecycle): wire cyclic perennation restart & REQ-047↔REQ-003 coupling (#565) #601 left these on generic sequences; Phase 3 feat(lifecycle): master-data sweep — dormancy models & herb/strawberry gaps (#565) #608 did the dormancy sweep)
• feat(lifecycle): D10 clonal continuation for monocarpic mothers #390 (D10 clonal continuation precedent)

Tracked from the Epic #565 closure — non-blocking optional refinement.

[nolte]

Resolved via #670 (merged to develop). Auto-close did not fire because the automerge app merges with an app token; closing manually.