multi-ai-code-review

Multi-perspective code review using Claude, Gemini, and Codex as specialized agents. 5-dimensional analysis (security, performance, maintainability, correctness, style) with LLM-as-judge consensus, quality scoring, and CI/CD integration. Use when reviewing PRs, auditing code quality, preparing production releases, or establishing code review workflows.

Quality

66%

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/multi-ai-code-review/SKILL.md

Quality

Content

50%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 body is action-dense with usable prompt templates and commands, but it is padded with Claude-known concept restatements, lacks enforced validation feedback loops, and ships a broken reference plus inlined content that should be split out. Actionability is the clear strength; conciseness and structure are the weaknesses.

Suggestions

Trim padded sections (Key Principles, Quality Targets, Anti-Patterns to Detect, Example Review Session) that restate OWASP/SQL-injection/XSS concepts Claude already knows, keeping only operation-specific guidance.

Replace optional validation with enforced validate→fix→retry checkpoints in each operation (e.g. "if Gemini flags false positives, revise findings and re-verify before reporting").

Create the missing references/performance-patterns.md and move the inlined CI/CD YAML and anti-patterns catalog into reference files so SKILL.md stays a lean overview with one-level-deep, real links.

Dimension	Reasoning	Score
Conciseness	The ~530-line body contains several padded sections that restate concepts Claude already knows — "Key Principles (validated by tri-AI research)", "Quality Targets", the 10-item "Anti-Patterns to Detect", and an illustrative "Example Review Session" — making it noticeably verbose rather than merely improvable.	2 / 5
Actionability	Concrete bash commands ("git diff main...HEAD > /tmp/pr_diff.txt", "gemini -p \"...\"") and a full GitHub Actions YAML block provide mostly executable guidance, with only minor gaps from the template placeholders ([CODE], [DIFF]) that must be filled in.	4 / 5
Workflow Clarity	Operations are sequenced with numbered steps, but validation checkpoints are marked optional ("Validate with Gemini (optional)") rather than enforced, and no explicit validate→fix→retry feedback loop is specified.	3 / 5
Progressive Disclosure	Sectioning exists and a References list is signaled, but a referenced file (references/performance-patterns.md) does not exist, and large blocks that belong in separate files (full CI/CD YAML, anti-patterns catalog, example session) are inlined into the SKILL.md body.	3 / 5
	Total	12 / 20 Passed

Description

83%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 is strong: third-person, concise, and clearly answers both what the skill does and when to use it with natural trigger phrases. Minor gaps in keyword synonyms and overlap with generic review skills keep it just below perfect on two dimensions.

Dimension	Reasoning	Score
Specificity	Lists several concrete capabilities — "5-dimensional analysis", "LLM-as-judge consensus", "quality scoring", and "CI/CD integration" — but stops just short of the comprehensive, fully-enumerated coverage of a 5.	4 / 5
Completeness	Explicitly states both what it does (multi-perspective review with 5-dimensional analysis, consensus, scoring, CI/CD) and when to use it ("Use when reviewing PRs, auditing code quality, preparing production releases, or establishing code review workflows") with concrete trigger phrases.	5 / 5
Trigger Term Quality	Natural trigger phrases like "reviewing PRs", "auditing code quality", and "preparing production releases" give good keyword coverage, though a few synonyms (e.g. "merge requests", "pull requests") are absent.	4 / 5
Distinctiveness Conflict Risk	The multi-AI ensemble niche is mostly distinct with clear triggers, but "reviewing PRs / auditing code quality" still overlaps with generic code-review skills, leaving minor conflict risk.	4 / 5
	Total	17 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (537 lines); consider splitting into references/ and linking	Warning
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
referenced_paths_exist	Referenced path issues: 1 missing	Warning

	Total	13 / 16 Passed

Repository: fernandezbaptiste/Skrillz
Path: skills/multi-ai-code-review/SKILL.md
Commit: 93ed392

Reviewed: 1 day 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.