What problem does it solve? A Thinkwise Software Factory repository commonly hosts multiple application models with multiple branches each, and most entities are keyed by model_id/branch_id. Queries without explicit scoping can silently match rows from the wrong model or branch, and staged writes can silently drop or cross-contaminate field values. This Skill pins down the correct connector, model, and branch once per task and documents verified write-API hazards so mutations land correctly. ## Core Features & Use Cases - Context resolution workflow: Determines which MCP connector (sf_mcp, meta_dev, indicium), which application model (model_id), and which branch (branch_id) apply, asking the user only when context is ambiguous. - Shared conventions for sibling skills: Enforces confirm-before-mutate planning, ask-don't-default design decisions, and live-sample verification of unfamiliar entity field names across all thinkwise_software_factory_* skills. - Write-hazard mitigations: Documents verified failure modes of the stage/patch/commit flow — silently dropped patch properties, cross-contamination from parallel writes, transient commit not-found errors, weak-entity key requirements, and truncated large text reads — with tested workarounds such as property ordering and OData substring paging. - Use Case: Before creating a control procedure via an MCP connector, run this base check to pin the connector, model, and branch, then carry those values into every subsequent execute_odata_query and stage_resource call. ## Quick Start Before making any Software Factory MCP calls, confirm which connector, application model, and branch this task targets.
