[v1 readiness][legolas] Parser and lockfile regression fixture gate

[JeremyDev87]

[v1 readiness][legolas] Parser and lockfile regression fixture gate

Goal

Back v1 bundle-risk findings with fixtures for stats files, lockfiles, duplicate packages, and lazy-load guidance so risk estimates are reproducible.

Source of Truth

• Parent epic: [v1 readiness] Legolas release gate epic (no version bump) #109
• Repo: JeremyDev87/legolas
• Parent implementation plan artifact: /Users/pjw/.hermes/thread-workcells/discord/282499436049858561/1508085811139051580/TASKS/v1-ddplan-issues/IMPLEMENTATION_PLAN.md
• Version bump, tag, publish, CI dispatch, branch, commit, and PR creation are out of scope for this issue.

Wiki Context Manifest

• Queries attempted: frontend evidence gates release evidence contract, code is not cheap AI maintenance cost ratio, alex core invariants SSoT No Silent Fallback idempotency, 12-factor agents runtime contract event log release workflow, overeager coding agents task scope destructive action, prompt assets skill issue planning evidence
• Wiki sources read: [wiki: frontend/evidence-gates.md], [wiki: principles/code-is-not-cheap.md], [wiki: principles/alex-core-invariants.md], [wiki: harness-engineering/12-factor-agents-runtime-contract.md], [wiki: harness-engineering/overeager-coding-agents-scope-contract.md], [wiki: harness-engineering/prompt-assets.md], [wiki: raw/2026-05-24-ai-maintenance-cost-ratio-source-map.md]
• Relevant wiki facts: v1 readiness requires artifact/runtime/evidence bundles, not only test-green; release and wrapper flows must preserve SSoT, atomicity, idempotency, and No Silent Fallback; destructive or fix-like commands need task-scope/evidence gates; AI-generated velocity must lower future maintenance cost rather than add broad v1 scope.
• remember-wiki applicability / skip reason: skipped; target repos are JeremyDev87/* personal CLI repos, not Remember/dramancompany repos.
• Non-wiki inference: repo-specific split below is based on live local repo docs/config/workflows and npm registry checks performed before issue creation.

Quality Lens Router Output

• Applicable gate families:
  • evidence-contract: v1 readiness, package install, CLI output, release/publish, and fallback claims need concrete proof.
  • simplicity-deletability: v1 work must prefer contract/evidence gaps over broad new features or nonessential abstractions.
  • tdd-systematic-debugging: false-positive, fallback, exit-code, and packaging regressions should be fixture/test-backed.
• Skipped gates:
  • frontend-design: not-applicable; CLI/runtime/docs work only.
  • vercel-agent-skills: not-applicable; no React/Next/Vercel/mobile surface.
  • deploy-token-safety: limited; no token/deploy mutation in this issue, only release credential evidence requirements.

Evidence Contract

• Required evidence:
  • Candidate SHA/version alignment evidence when a candidate is selected.
  • Repo test evidence plus source CLI smoke.
  • npm pack / packed-install evidence for the user-consumed artifact.
  • Release workflow/check/dist-tag/GitHub Release evidence where release automation is touched.
  • Explicit fallback/unsupported-platform evidence where wrapper/native-runtime selection is touched.
• Evidence not applicable:
  • Browser screenshot/DOM: not-applicable, CLI repos.
  • Production URL: not-applicable, npm/GitHub release surface only.
• Blocking evidence gaps:
  • Final v1 version bump/tag is intentionally out of scope for this issue set.

Owned files / likely touched areas

• crates/legolas-core/src/**
• tests/**
• fixtures/**
• crates/legolas-cli/tests/**

Inspect-only files

• README*.md
• docs/plan/**

Must not touch

• claim exact byte savings where only risk signal exists
• broad optimizer rewrites outside fixture evidence

Implementation steps

1. Inventory supported stats/lockfile inputs.
2. Add fixtures for npm/pnpm/yarn and common bundler stats if not covered.
3. Ensure findings include origin/evidence path and caveat text.
4. Add edge-case regression tests for duplicate and lazy-load candidates.

Parallelization note

Can run independently from packaging docs if fixture paths do not overlap.

Dependencies / blocked by

Coordinate with L2 if output fields change.

Validation / evidence required

• cargo test --workspace
• npm run smoke
• targeted parser/fixture tests

Completion criteria

• Required validation evidence is pasted into the PR body or final issue comment.
• No v1 version bump/tag/publish is performed.
• If any required evidence is unavailable, the issue remains open with the gap classified as Medium/High rather than being marked complete.