vm-incremental-planner
Skill detail with category, linked agents, and source metadata.
vm-incremental-planner
Inventory existing RE artifacts and produce a KEEP/GAP-FILL/REFRESH plan per phase. Enables incremental re-runs that preserve completed work and focus only on gaps. Use when run_mode=incremental or resuming a partial RE run. Writes intermediate/incremental-plan.json.
Source: .github/skills/compliance/vm-incremental-planner/SKILL.md
Used By Agents
Preview
View source preview (first 3000 chars)
# Skill: VM Incremental Planner ## Purpose Inventory existing RE artifacts and decide per-phase whether to KEEP, GAP-FILL, or REFRESH. Produces an incremental plan that lets the RE orchestrator skip completed work and focus only on gaps. ## When to Invoke - Before any RE run when `run_mode=incremental` - After a partial RE run that was interrupted - When the user explicitly asks to resume or continue a previous RE run ## Inputs | Input | Source | Required | |---|---|---| | `output_dir` | RE output directory (e.g., `docs/codebase-analysis`) | Yes | | `checklist_path` | Path to `checklist.json` | Yes | ## Decision Logic ### Phase 0 (Inventory) - **KEEP** if `phase0-inventory.json` exists and has > 0 files - **REFRESH** if missing ### Phase -1 (Script Calibration) - **KEEP** if `script-config.json` exists and has > 0 language entries - **REFRESH** if missing or empty ### Phase 1 (Discovery) - **KEEP** if `discovery.json` exists with non-empty frameworks + dependencies - **REFRESH** if missing ### Phase 2 (Signatures) - **KEEP** if `signatures.json` exists and entry count >= 90% of inventory file count - **GAP-FILL** if exists but < 90% coverage - **REFRESH** if missing ### Phase 3 (Dependency Graph) - **KEEP** if `dependency-graph.json` exists with edges > 0 - **REFRESH** if missing ### Phase 4a (File Metrics) - **KEEP** if `file-metrics.json` entry count >= 95% of inventory - **GAP-FILL** if exists but < 95% - **REFRESH** if missing ### Phase 4b (File Analysis / AI Summaries) - **GAP-FILL** if `file-analysis.json` exists but hollow% > 5% - **KEEP** if hollow% <= 5% - **REFRESH** if missing ### Phase 5 (Data Model) - **KEEP** if `data-model.json` exists with > 0 entities - **REFRESH** if missing ### Phase 6 (User Flows) - **KEEP** if `user-flows.json` exists with > 0 flows - **REFRESH** if missing ### Phase 7 (Architecture) - **KEEP** if `architecture.json` exists with > 0 layers - **REFRESH** if missing ### Phase 8 (Business Logic) - **KEEP** if `business-logic.json` and `business-rules.json` both exist with > 0 entries - **REFRESH** if either missing ### Phase 9 (Cross-Linking) - Always **REFRESH** in incremental mode (cheap, ensures consistency after gap-fills) ### Phase 10 (Reports) - Always **REFRESH** in incremental mode (ensures HTML reflects updated data) ### Phase 11 (Brownfield) - **KEEP** if `brownfield-context.json` exists with > 0 conventions AND `feature_flags` key is present - **GAP-FILL** if exists with conventions but `feature_flags` key is absent (pre-v5.2.0 artifact) - **REFRESH** if missing ### Phase 10D (Ops Context) [v5.2.0] - **KEEP** if `ops-context.json` exists with non-empty `config_registry` AND non-empty `startup_sequence` - **REFRESH** if missing or incomplete ### Phase 10E (Key Findings) [v5.3.0] - **KEEP** if `key-findings.json` exists with > 0 total findings (functional + nonfunctional + business_flow) AND > 0 prioritised_recommendations - **REFRESH** if missing or empty ### Enriched Sub-Agent O