What problem does it solve? When two AI agents working in parallel produce conflicting git branches, neither agent can fairly resolve the conflict — each lacks the peer's context and is biased toward its own changes, leading to overwritten work or abandoned features. This Skill provides an impartial reconciliation procedure that merges both sides based on their stated intents. ## Core Features & Use Cases - Hunk Classification: Categorizes every conflicted hunk as disjoint-intent, same-question-different-answer, or superseded, then applies the correct resolution strategy per hunk. - Impartiality Contract: Enforces strict rules — never favor the spawning side, touch only conflicted regions, and surface every design decision explicitly in the hand-back summary. - Kanban-Native Workflow: Integrates with hermes kanban cards so a third neutral profile can be assigned reconciliation tasks with both conflicted cards linked as parents. - Use Case: During a parallel-PR wave, two agents modify the same module and the merge halts. Spawn a neutral subagent with both branch names and intent summaries; it classifies each hunk, resolves conflicts, runs the build/tests, commits the merge, and reports every decision for human veto. ## Quick Start Ask the agent to act as a neutral reconciler for the halted merge in this repository, using the completion summaries from both conflicting branches, and follow the merge-reconciler procedure to produce a verified merge commit.
