Merge Conflict Resolver

The Problem

You merge a feature branch and hit 14 conflicts across 6 files. The agent picks "ours" for all of them because that's the fastest resolution. It compiles. It passes tests. You deploy. Two weeks later, a customer reports that a critical database migration column is missing — it was on the feature branch side. The "fast resolution" silently dropped a migration, and no one caught it because the tests didn't cover the schema change. Merge conflict resolution isn't a speed contest — it's a correctness exercise with a specific methodology.

What You Get

• Classify every conflict by type — content, add/add, rename/rename, rename/delete, modify/delete, binary, or submodule — so you apply the correct resolution strategy for each
• Map conflict causes to resolution strategies — divergent function edits (manual merge), whitespace reformatting vs. logic changes (accept reformat, re-apply logic), API contract divergence (update all call sites)
• Provide language-specific resolution guidance — Python indentation consistency, TypeScript import deduplication, Go sorted imports, Java balanced braces, SQL migration number conflicts and foreign key preservation
• Verify resolution integrity — check for remaining conflict markers, run build/compile, execute test suite, verify import consistency, and confirm no accidental deletions of unrelated code
• Handle binary file conflicts — extract both versions to temp files, compare content, and decide which to keep or create a new merged version
• Prevent future conflicts — recommend merge frequency, CODEOWNERS setup, short-lived branches, feature flags, and consistent auto-formatting

Why This Beats Prompting It Yourself

Telling an agent "resolve the conflicts" produces one of two outcomes: it picks one side blindly (losing the other's changes) or it concatenates both sides (creating duplicate logic). This skill runs a 9-step workflow: assess scope, understand cause, select strategy, resolve by conflict type, handle language-specific pitfalls, verify with build + tests, and complete with a clean commit. The verification step catches the silent failures that blind resolution misses.

Example Session

You: "Merge feature/auth-v2 into main — I have conflicts in 3 files"

Agent:

══════════════════════════

Conflicted files: 3
Conflict types: 2 content, 1 dependency version

FILE 1: SRC/AUTH/LOGIN.TS (CONTENT CONFLICT)
────────────────────────────────────────────
Cause: Both branches modified validateToken() — ours added rate limiting, theirs added MFA
Resolution: Manual merge — both changes are needed. Combined rate limiting + MFA validation.
Verification: Import statements consistent ✓, function signature unchanged ✓

FILE 2: SRC/AUTH/TYPES.TS (CONTENT CONFLICT)
────────────────────────────────────────────
Cause: Our branch added TokenPayload interface, theirs added MFAChallenge
Resolution: Accept both — non-overlapping additions.

FILE 3: PACKAGE.JSON (DEPENDENCY CONFLICT)
──────────────────────────────────────────
Cause: Both branches updated jsonwebtoken — ours to ^9.0.1, theirs to ^9.0.2
Resolution: Accept ^9.0.2 (newer), run npm audit ✓, verify no breaking changes ✓

VERIFICATION
────────────
- [x] Zero conflict markers remaining
- [x] Build pas

Use Cases

• Resolving multi-file merge conflicts after long-lived feature branches
• Handling rename/delete conflicts during module restructuring projects
• Merging SQL migration files where both branches added schema changes
• Reviewing and correcting blind "accept ours/theirs" resolutions
• Onboarding junior devs — teach conflict anatomy and resolution strategy
• Setting up team conventions for preventing future conflicts (CODEOWNERS, short branches)

Known Limitations

Binary file conflicts require manual inspection — the skill can extract and compare but cannot merge binary content. Submodule conflicts depend on the upstream submodule state. Complex three-way merge conflicts in auto-generated files (protobuf, OpenAPI specs) may require regenerating the file rather than manual merge.

Upgrade to Pro

Free resolves the standard conflicts. Merge Conflict Resolver Pro ($5) handles the hard cases: rename/delete, binary, submodule, and semantic conflicts that compile but break. Upgrade for long-lived branches or daily conflict traffic — Pro version.