Rubric: L-PREFLIGHT (Mechanical inventory)
Lens: L-PREFLIGHT · Density: D-HIGH · Axes: CMP, MOD
Purpose
Establish what exists before opinions form: languages, module graph sketch, build/test entrypoints, generated vs hand-written code, obvious binary/secrets risk paths.
Evaluation criteria (exhaustive within scope)
- Detect languages/build systems (go.mod, package.json, Cargo.toml, pom.xml, etc.).
- Enumerate top-level modules and approximate LOC by area.
- Identify test runners and CI config presence/absence.
- Flag generated trees (
*.pb.go,node_modules,dist/,vendor/). - List available analyzers (linters, formatters, security scanners, AST tools).
- Produce initial C4-L1/L2 diagram drafts with anchors (may be refined in L-ARCHITECTURE).
Evidence expectations
Command outputs, directory listings, tool version strings. Prefer E2.
Citations
| Work | Point |
|---|---|
| ISO/IEC 25010 | Product quality — need inventory before characteristic scoring |
| Software engineering practice | Baseline metrics before qualitative review (common SEI/measurement guidance) |
Pros / cons
| Pros | Cons |
|---|---|
| Cuts hallucination by grounding later lenses | Inventory alone is not quality |
| Surfaces tooling gaps early | LOC is a weak proxy — label as approximate |
Deliverables
preflight.json, draft diagrams/D-CONTEXT-01.md, diagrams/D-CONTAINER-01.md, tooling_gaps.md seed.