chore(types): remove dead protoMethodsMs field and stale comment

[carlos-alm]

The core work for #1432 landed in b5c03a2 — runPostNativePrototypeMethods was deleted and Pattern 3 (fn.method = function(){} / arrow) was added to the Rust extractor. Two dead-code artifacts were not cleaned up:

• protoMethodsMs?: number in BuildResult.phases — declared, never set by any surviving code path
• The parseFilesWasmForBackfill docstring still mentioned "prototype-methods post-pass" alongside this-dispatch, but only the this-dispatch post-pass uses symbolsOnly now

Removes both. Lint and prototype-method-resolution integration tests (WASM + native) pass.

Closes #1432

[claude]

Claude finished @carlos-alm's task —— View job

---

PR Review

• Gather context
• Read changed files
• Analyze phases 1-5
• Post final assessment

---

Phase 1: Root Cause Analysis

Problem Being Solved: Two dead-code artifacts remained after runPostNativePrototypeMethods was deleted in b5c03a2 (#1432):

1. protoMethodsMs?: number in BuildResult.phases — declared but never written by any surviving code path
2. The parseFilesWasmForBackfill JSDoc mentioned "prototype-methods and this-dispatch post-passes" when only the this-dispatch post-pass remains

Why This Approach: These are straightforward deletions of genuinely dead code — no behavior changes, no logic alterations. Correct approach.

Risk Assessment: Minimal. protoMethodsMs was optional (?:), so any code reading it already had to handle undefined. Removing it from the type contract makes the interface more honest. The docstring fix is a pure comment edit.

---

Phase 2: Configuration Change Detection

No lint rules changed, no tsconfig modifications, no test configuration touched. Clean.

---

Phase 3: Test Integrity

No test files modified. The PR description states "prototype-method-resolution integration tests (WASM + native) pass" — consistent with a pure dead-code removal that cannot affect runtime behavior.

---

Phase 4: Code Quality

src/types.ts:1268–1270 — Removal of protoMethodsMs?: number is correct. The field was optional, so removing it is a non-breaking change at runtime (existing callers reading .phases.protoMethodsMs will just get undefined as before — no change). At the type level it is now properly absent from the contract, which is strictly better.

src/domain/parser.ts:1262–1264 — The docstring update is accurate. Before: "prototype-methods and this-dispatch post-passes". After: "this-dispatch post-pass". This reflects the current code reality: parseFilesWasmForBackfill is called from the this-dispatch post-pass with symbolsOnly: true, and from the dropped-language backfill with symbolsOnly: false (default). The prototype-methods pass no longer exists. The edit is minimal and precise.

No concerns on either change.

---

Phase 5: Backlog Compliance

• Zero-dep: ✓ — No dependency changes whatsoever.
• Foundation-aligned: ✓ — Pure cleanup; no new infrastructure, no cloud calls, no agent behavior.
• Problem-fit: 2/5 — Maintenance cleanup. Doesn't directly reduce token waste or prevent hallucinated deps, but type hygiene marginally improves agent accuracy when consuming BuildResult. Appropriate to bundle with the parent fix (Native engine: extract func-prop methods (fn.method = function(){}) in Rust to eliminate the WASM post-pass #1432).
• Breaking: No — protoMethodsMs was ?: optional; removing it from the type is backward-compatible at the value level. Any consumer that compiled with protoMethodsMs: someValue would have been a type error before this PR anyway, since nothing set it.
• Tier: 1 (zero-dep ✓, foundation-aligned ✓)

---

Final Assessment

## Root Cause Analysis
Problem Being Solved: Remove two dead-code artifacts left over from the #1432 deletion of
  runPostNativePrototypeMethods.
Why This Approach: Smallest correct fix — delete the optional field from the type contract and
  update the docstring to match the surviving code path.
Risk Assessment: Negligible. The field was optional and never written; callers already handled
  undefined. No runtime behavior changes.

## Backlog Compliance
- Zero-dep: ✓
- Foundation-aligned: ✓
- Problem-fit: 2/5 — maintenance cleanup, not a user-facing feature
- Breaking: No
- Tier: 1

## Critical Concerns
None.

## Final Recommendation
- Rating: ⭐⭐⭐⭐ (4/5)
- Action: APPROVE
- Reasoning: Correct, minimal, well-scoped cleanup. The two removals are independently
  verifiable as dead code. Commit message accurately describes both changes and links the
  parent issue. No configuration relaxations, no test weakening, no new deps.

[greptile-apps]

Greptile Summary

This PR cleans up two dead-code artifacts left over after runPostNativePrototypeMethods was deleted in b5c03a2: the protoMethodsMs?: number field in BuildResult.phases (never set by any surviving code path) and the stale "prototype-methods post-pass" mention in the parseFilesWasmForBackfill docstring. It also introduces a new /resolve skill for context-aware merge-conflict resolution — an addition not mentioned in the PR description.

• src/types.ts (visible in commit 61a9839, net-neutral in the final diff vs main): removes the protoMethodsMs optional field and its JSDoc.
• src/domain/parser.ts: tightens the parseFilesWasmForBackfill docstring to reference only the surviving this-dispatch post-pass.
• .claude/skills/resolve/SKILL.md: adds a 527-line, 7-phase skill that checks out a PR branch, traces conflict origins to their source PRs via gh, resolves with full context, verifies both-sides intent, and pushes — with an explicit stop condition for ambiguous conflicts.

Confidence Score: 5/5

Safe to merge: the TypeScript changes are pure dead-code and docstring cleanup with no logic impact; the new skill file operates only when explicitly invoked and touches no production code paths.

Both TypeScript changes are mechanical removals of a field that was never written and a comment that was no longer accurate — neither alters any runtime behaviour. The new SKILL.md adds a developer tool with no production coupling. The only real nit is that the skill hardcodes optave/codegraph rather than deriving the repo dynamically, but this is a maintenance concern, not a correctness defect in the current setup.

.claude/skills/resolve/SKILL.md — the hardcoded --repo optave/codegraph in every gh call is worth revisiting before the skill is shared or the target repo is renamed.

Important Files Changed

Filename	Overview
.claude/skills/resolve/SKILL.md	New /resolve skill for context-aware merge-conflict resolution across 7 phases; hardcodes optave/codegraph throughout, which works for this ops-tool's intended target but would silently fail against any other repo without edits. PR description omits this addition entirely.
src/domain/parser.ts	Docstring-only update: removes stale "prototype-methods post-pass" reference from parseFilesWasmForBackfill; no logic changes.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A["/resolve <pr-number>"] --> B["Phase 0: Pre-flight\nVerify git, gh, jq; fetch PR metadata"]
    B --> C["Phase 1: Branch Setup\ngh pr checkout; git merge origin/base"]
    C --> D{Conflicts?}
    D -- No --> E["Exit clean"]
    D -- Yes --> F["Phase 2: Conflict Archaeology\nTrace incoming commits → source PRs\nFetch source PR descriptions & diffs"]
    F --> G["Phase 3: Current PR Context\nRead this PR's description & per-file diffs"]
    G --> H["Phase 4: Resolution\nEdit files; apply both-sides intent\nStage by name (never git add .)"]
    H --> I{Ambiguous?}
    I -- Yes --> J["STOP — explain to user\nwhat decision is needed"]
    I -- No --> K["Phase 5: Verification\nDiff vs both parents; confirm\nno intentional change was dropped"]
    K --> L["Phase 6: Validation\nRun tests & lint"]
    L --> M["Phase 7: Commit & Push\ngit commit; git push\nclean up .codegraph/resolve/"]

Loading

_{Reviews (2): Last reviewed commit: "fix: resolve merge conflicts with main" | Re-trigger Greptile}

[github-actions]

Codegraph Impact Analysis

1 functions changed → 0 callers affected across 0 files

• BuildResult.phases in src/types.ts:1261 (0 transitive callers)