User Story
I develop OpenShell with OpenCode and use the repository's principal-engineer-reviewer sub-agent to review my changes. The OpenCode copy gives different, older guidance than the Claude Code copy and gator.
Problem Statement
The reviewer persona is kept in two hand-maintained copies: .claude/agents/principal-engineer-reviewer.md and .opencode/agents/principal-engineer-reviewer.md. Recent reviewer changes landed only in the Claude copy, so the OpenCode copy is missing the pragmatic review calibration, the evidence requirements, and the current severity definitions.
Impact / Why This Matters
Contributors who use OpenCode get a reviewer that is more prone to scope creep and doesn't match what gator enforces on PRs. Any future change to the reviewer has to be copied by hand, and the copies drift again.
Acceptance Criteria
Reproduction Steps
- Compare
.claude/agents/principal-engineer-reviewer.md with .opencode/agents/principal-engineer-reviewer.md on main.
- Observe that the OpenCode copy has no "Pragmatic review calibration" section.
User Story
I develop OpenShell with OpenCode and use the repository's
principal-engineer-reviewersub-agent to review my changes. The OpenCode copy gives different, older guidance than the Claude Code copy and gator.Problem Statement
The reviewer persona is kept in two hand-maintained copies:
.claude/agents/principal-engineer-reviewer.mdand.opencode/agents/principal-engineer-reviewer.md. Recent reviewer changes landed only in the Claude copy, so the OpenCode copy is missing the pragmatic review calibration, the evidence requirements, and the current severity definitions.Impact / Why This Matters
Contributors who use OpenCode get a reviewer that is more prone to scope creep and doesn't match what gator enforces on PRs. Any future change to the reviewer has to be copied by hand, and the copies drift again.
Acceptance Criteria
Reproduction Steps
.claude/agents/principal-engineer-reviewer.mdwith.opencode/agents/principal-engineer-reviewer.mdonmain.