migration

name	migration
description	Convert staged legacy source files in `migrate/inputs/` into LN2 inventory YAML at `migrate/output/ln2_inventory.yaml`, validate the result, and import it as a new managed dataset. Use when the user is running migration onboarding or asks to transform external source materials into LN2 inventory format.

Migration

Follow this skill when handling staged migration inputs.

Core Workflow

1. Inspect source files staged in migrate/inputs/.
2. Keep live progress in migrate/output/migration_checklist.md.
3. Propose the field mapping and schema plan before conversion.
4. Ask for explicit approval before locking migrate/output/expected_schema.json.
5. Materialize final output at migrate/output/ln2_inventory.yaml.
6. Run validate with path set to migrate/output/ln2_inventory.yaml before any import.
7. Import only after collecting target_dataset_name and explicit CONFIRM_IMPORT.

Read Next As Needed

• Workflow overview: references/workflow.md
• Prompt framing: references/prompt_en.md
• Detailed runbook: references/runbook_en.md
• Shared structural fields: ../shared/references/structural_fields.md
• Shared schema context: ../shared/references/schema_context.md
• Shared validation contract: ../shared/references/validation_contract.md
• Shared audit and confirmation policy: ../shared/references/audit_policy.md

Non-Negotiables

• Do not invent records, dates, positions, or metadata.
• Do not bypass validate(path="migrate/output/ln2_inventory.yaml").
• Do not import before explicit human confirmation.
• When source YAML is already a valid LN2 document and mapping is identity, use fs_copy for whole-file passthrough instead of re-emitting file contents with fs_write.
• Resume from existing migrate/output/expected_schema.json, migrate/output/migration_checklist.md, and migrate/output/ln2_inventory.yaml when they are already present and still valid.