Files
info-tech-canon/infospace/agent/review-kit/review-workflow.yaml

101 lines
3.9 KiB
YAML

id: review-kit/alignment/workflow
title: Canon Consumer Alignment Review Workflow
status: candidate
created_by_workplan: ITC-WP-0011
workflow_goal: Turn an unknown or partially understood repository into a source-linked canon alignment assessment and repo-local workplan proposal.
phases:
- id: intake
title: Repository Intake
purpose: Capture what the repository appears to provide before selecting canon surfaces.
inputs:
- README, INTENT, SCOPE, architecture docs, package manifests, tests, API specs, deployment files, and existing workplans.
questions:
- What producer intent does the repository declare?
- What current scope is evidenced by code, docs, tests, and deployment artifacts?
- Who are the likely consumers and what purposes do they bring?
- Which parts are owned by the repository and which are dependencies, adapters, generated assets, or examples?
outputs:
- repository_context
- initial_intent_scope_purpose_notes
- evidence_inventory
- id: surface-selection
title: Canon Surface Selection
purpose: Choose the smallest useful set of models, standards, profiles, benchmarks, and evaluation packs for the review.
inputs:
- repository_context
- model_selection_guide
- retrieval_index
questions:
- Which domain models are directly relevant?
- Which standards or benchmarks should stress-test the review?
- Which existing consumer packs already match the repository shape?
outputs:
- selected_canon_surfaces
- surface_selection_rationale
- id: mapping
title: Repository To Canon Mapping
purpose: Map observed repository concepts, entities, edges, APIs, policies, tasks, and artifacts to canon concepts.
inputs:
- selected_canon_surfaces
- evidence_inventory
questions:
- Which repository concepts map directly?
- Which concepts are partial fits or conflicts?
- Which mappings require stronger evidence?
- Which native terms must be preserved to avoid semantic flattening?
outputs:
- mapping_findings
- evidence_links
- unresolved_questions
- id: scoring
title: Fit Scorecard
purpose: Score alignment consistently enough to compare reviews and prioritize follow-up.
inputs:
- mapping_findings
- scorecard
questions:
- Where is fit strong enough to rely on?
- Which gaps block integration or reuse?
- Which conflicts risk wrong interpretation?
- Which findings are urgent, useful, or merely interesting?
outputs:
- scorecard_results
- prioritized_gaps
- risk_notes
- id: workplan-proposal
title: Consumer Workplan Proposal
purpose: Convert findings into repo-local workplans without moving consumer implementation work into the canon repo.
inputs:
- scorecard_results
- consumer_workplan_template
questions:
- What should the consumer repo change first?
- What validation should prove the alignment improved?
- Which tasks are refactors, documentation, schema changes, data migration, or governance review?
outputs:
- recommended_workplans
- proposed_tasks
- validation_plan
- id: canon-feedback
title: Canon Feedback And Pressure
purpose: Capture where the review suggests canon improvements without altering models or standards silently.
inputs:
- mapping_findings
- unresolved_questions
- scorecard_results
questions:
- Which gaps indicate the canon needs a new concept, profile, pattern, schema, or benchmark?
- Which findings should stay consumer-specific?
- Which proposed canon changes need a separate workplan?
outputs:
- canon_feedback
- proposed_canon_workplans
- open_questions
minimum_review_packet:
- repository_context
- selected_canon_surfaces
- mapping_findings
- scorecard_results
- recommended_workplans
- canon_feedback