Open Code Review Delegated Review
Let the host agent perform the review while OCR provides deterministic file selection and rule resolution.
The skill explicitly delegates the substantive review to the host agent and limits OCR to deterministic file selection and rule resolution, which is relatively restrained; however, it does not document sensitive-code, credential, or rule-content data flows, user confirmation, external side effects, rollback, or output isolation. Global npm installation is suggested without dependency-integrity or version-pinning guidance, so trust is not scored highly.
The workflow, three review modes, untracked-file handling, and output fields are broadly consistent; however, there are no tests for the delegate preview/rule paths, no pinned CLI version, no abnormal-input behavior, and no diagnosable failure guidance. Repository CI provides generic Go test and build evidence but does not reproduce this skill path, so the static ceiling applies.
The audience and scenario are clear: a host AI agent performs the review while OCR handles deterministic file selection and rule resolution; inputs, commands, and comment fields are also reasonably explicit. Boundaries for non-Git repositories, very large changes, missing rules, or insufficient host-agent capability are under-specified. The documentation is English-only and depends on npm/GitHub reachability, with no evidence of stable mainland-China network fit.
SKILL.md includes a name, description, Apache-2.0 license, author, homepage, and version, and is organized around prerequisites, workflow, formatting, classification, and gotchas. Shared material adds the license, security policy, CI, and installation guidance. It loses points for lacking a changelog, explicit maintenance ownership or update path, and for leaving the CLI dependency unpinned.
The skill connects file selection, rule retrieval, diff acquisition, review, and comment formatting into a directly usable workflow, and the fields support structured comments. However, the substantive review is entirely dependent on the host agent; the skill supplies no verified review result and leaves context handling, judgment, and failure recovery to the host. Static evidence therefore supports only limited core effectiveness.
The commands and workflow are auditable from the source, while repository CI, Govulncheck, and the coverage threshold provide limited corroboration. There are no delegate-specific tests, committed representative outputs, or third-party reproduction evidence. README claims about scale, benchmarks, and effectiveness cannot be independently verified from the supplied files, so the score remains near the static ceiling but below full marks.
- Review quality, line accuracy, and rule applicability depend on the host agent and are not guaranteed by OCR itself.
- Before installation, inspect the npm package, remote installer, and binary provenance, and pin a verifiable version or commit; the skill lacks confirmation, sensitive-data redaction, and outbound-scope controls.
- In mainland-China restricted networks, npm/GitHub downloads may be unreachable; prepare a local, mirrored, or verified binary in advance.
- Before reading untracked files or invoking Git commands, confirm the repository path, review scope, and that file contents will not cause unintended data exposure.
What it does & when to use it
This skill enables Open Code Review’s delegation mode. OCR handles deterministic engineering tasks: identifying reviewable files and resolving the rules that apply to them. The host agent performs the actual code review with its own model capabilities and tools. OCR does not need an LLM endpoint configured in this mode, making it suitable when review decisions should remain under the host agent’s control.
Checks whether the ocr CLI is installed; runs ocr delegate preview to obtain the workspace, range, or commit mode, reference metadata, reviewable files, and exclusions; runs ocr delegate rule <path...> to retrieve grouped rules; uses git diff, git show, or direct file reads according to the preview mode; reviews each file using its diff, rule group, and relevant context; and produces structured findings with a path, comment content, optional new-file line numbers, category, and severity.
- A team wants Claude Code or Codex to make the review judgments while OCR keeps file selection deterministic.
- A developer needs to review a branch range using preview-provided merge-base and reference metadata.
- A maintainer wants line-targeted findings for a single commit.
- A team reviewing a large change wants to fetch rules in batches and avoid repeating identical rule content.
- An engineer needs controlled review of tracked, untracked, and excluded workspace files.
Pros & cons
- Uses deterministic CLI logic for file filtering and rule resolution.
- Keeps review reasoning, context retrieval, and comment generation with the host agent.
- Supports workspace, branch-range, and single-commit preview modes.
- Groups files sharing the same rule content to reduce repetition.
- Defines structured categories and severity levels for findings.
- Requires the `ocr` CLI and shell access.
- The source does not document a dedicated installation or registration command for this specific skill across clients.
- The skill itself does not perform the LLM review; quality depends on the host agent.
- No independent benchmark, test suite, or platform validation is provided for this delegation skill.
- Untracked workspace files must be read directly rather than obtained through a normal Git diff.
How to install
Install the OCR CLI with npm install -g @alibaba-group/open-code-review, or use a GitHub release binary. Make the skill available at skills/open-code-review-delegate/SKILL.md in a host that supports Agent Skills. The source does not document a separate client-specific installation command for this delegation skill.
How to use
From the target Git repository, run ocr delegate preview; optionally provide --from <ref> --to <ref>, --commit <hash>, or --exclude <patterns>. Pass the reviewable paths to ocr delegate rule <path1> <path2> .... For range mode, use git diff <merge_base>..<to> -- <path>; for commit mode, use git show <commit> -- <path>; for workspace mode, use git diff HEAD -- <path> or read an untracked file directly. Review each file with the host agent and report findings using path, content, optional line fields, category, and severity. No OCR-side LLM configuration is required.
Compared to similar skills
Compared with OCR’s mode in which OCR calls an LLM endpoint, delegation mode leaves the actual review to the host agent and requires no OCR-side LLM configuration; OCR remains responsible for file selection and rule resolution.