review-remote

name	review-remote
description	Review open PR/MR and publish findings.
model	sonnet
allowed-tools	["Read","Write","Edit","Glob","Grep","Bash","WebFetch","mcp__*"]

Run the reviewer agent in remote mode on the current branch's open PR/MR. Analyzes the diff against repository-local checklists, instructions, and built-in heuristics. Generates a consolidated review iteration YAML (`review-iter-.yaml`) locally; optionally publishes findings and records publish URLs in that same YAML. User invocation: /review-remote [options] Examples: /review-remote /review-remote --publish /review-remote --pr 42 /review-remote --gitlab --publish /review-remote --mr 15 --dry-run /review-remote --github --pr 100 --publish $ARGUMENTS --github | --gitlab: Force platform detection. OPTIONAL. --pr | --mr : Explicit PR/MR number. OPTIONAL (auto-detected from branch). --publish: Publish findings to remote (default: dry-run). OPTIONAL. --dry-run: Explicit dry-run mode (this is also the default). OPTIONAL. Parse $ARGUMENTS and delegate to the `reviewer` agent (remote mode) with the parsed options. The agent handles all platform detection, pre-flight checks, review, and optional publishing. This command is a thin entry point — all logic lives in the `reviewer` agent. Non-interactive: do not depend on follow-up questions; use $ARGUMENTS, safe defaults, or NEEDS_INPUT. Idempotent: reruns are safe; prior `review-iter-*.yaml` files and external comments prevent duplicate findings/comments.

<output_format> <what_to_return>Review findings summary (count, severity breakdown), the review-iter-<N>.yaml path under tmp/code-review/<branchPath>/, publish state (draft or published), and next action suggestion.</what_to_return> </output_format>

<user_input>$ARGUMENTS</user_input>