open-code-review-delegate

Delegation mode for open-code-review (OCR). Instead of OCR calling an LLM endpoint, this skill instructs the host agent to perform the code review itself, using OCR only for deterministic engineering: file selection and rule resolution. Use when the host agent should drive the review with its own LLM capabilities.

Quality

78%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./skills/open-code-review-delegate/SKILL.md

Quality

Content

88%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The content is a highly actionable, well-sequenced workflow with executable commands, explicit validation checkpoints, and error-recovery feedback loops. Its only notable weakness is that the longer reference and edge-case material is kept inline rather than offloaded to one-level-deep reference files.

Dimension	Reasoning	Score
Conciseness	The body is information-dense, uses tables and code blocks, and assumes Claude knows git and code review without padding, but a few passages (the shell-injection warning in the background-context section and the extended --format troubleshooting explanation) could be trimmed, fitting anchor 4 rather than 5 (every token earns its place) or 3 (includes unnecessary explanation).	4 / 5
Actionability	Concrete, copy-paste-ready commands are given throughout (ocr delegate preview/rule, git diff/show invocations) with a table of common cases (workspace/range/commit) and a fully specified output schema, matching anchor 5 (fully executable, covers common cases) rather than 4 (minor gaps).	5 / 5
Workflow Clarity	Seven clearly sequenced steps include a per-file checklist (Step 4), explicit coverage validation before reporting (Step 6), and feedback loops for error recovery (oversized-background retry and --format version troubleshooting), satisfying anchor 5; the batch-operation cap at 3 does not apply because validation/verification steps are explicitly present.	5 / 5
Progressive Disclosure	The body is well-organized with clear section headers (Workflow, Sub-commands Reference, Shared Flags, Gotchas) and no nested references, but all reference material and two detailed edge-case sections are inlined in a ~170-line document rather than split into one-level-deep files, fitting anchor 4 rather than 5 (clear overview pointing to well-signaled separate references).	4 / 5
	Total	18 / 20 Passed

Description

67%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description clearly states what the skill does and when to use it, with concrete actions and a distinct OCR-delegation niche. Its main weakness is that the "Use when" trigger is an architectural condition rather than a phrase a user would naturally say, limiting trigger-term quality.

Suggestions

Add user-facing trigger phrases to the 'Use when...' clause (e.g., 'Use when the user asks to review code, a PR, or staged changes with the OCR tool and wants the host agent to do the reviewing').

Include common synonyms/variations such as 'code review', 'PR review', 'diff review', and 'review changes' to improve natural keyword coverage.

Tighten the architectural framing so the trigger reads as a user need rather than an internal mode selection.

Dimension	Reasoning	Score
Specificity	Names concrete actions for each party — OCR does "file selection and rule resolution" while the host "perform[s] the code review itself" — giving several specific actions with only minor coverage gaps, fitting the anchor 4 rather than 3 (which requires only 1-2 actions) or 5 (which expects a comprehensive list).	4 / 5
Completeness	Both "what" (delegation mode where the host performs the review and OCR handles deterministic engineering) and "when" (explicit "Use when..." clause) are present, but the "when" is a system-level condition rather than a concrete user-facing trigger phrase, fitting anchor 4 rather than 5 (concrete trigger phrases) or 3 ("when" missing/weakly implied).	4 / 5
Trigger Term Quality	The natural keyword "code review" is present, but the trigger guidance ("Use when the host agent should drive the review with its own LLM capabilities") is an architectural condition a user would rarely voice, and common synonyms/variations are missing, matching anchor 3 rather than 4 (good keyword coverage) or 2 (only generic keywords).	3 / 5
Distinctiveness Conflict Risk	"Delegation mode for open-code-review (OCR)" carves a clear niche tied to a specific tool's mode, but the broad trigger (host agent driving a review) carries minor overlap risk with general code-review skills, fitting anchor 4 rather than 5 (minimal conflict risk) or 3 (could still overlap with similar skills).	4 / 5
	Total	15 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: alibaba/open-code-review
Path: skills/open-code-review-delegate/SKILL.md
Commit: 124bfc3

Reviewed: ~19 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.