agent-code-review-swarm

Agent skill for code-review-swarm - invoke with $agent-code-review-swarm

2.64x

Quality

33%

Does it follow best practices?

Impact

98%

2.64x

Average score across 3 eval scenarios

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./.agents/skills/agent-code-review-swarm/SKILL.md

Quality

Content

38%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 reads as a feature catalog for the ruv-swarm tool rather than a skill: command dumps with unverifiable flags, duplicated sections, a stray second metadata block, corrupted file paths, and dangling references to nonexistent bundle files. It needs consolidation into a short overview with a real end-to-end workflow, validation steps, and working links to actual reference files.

Suggestions

Replace the command catalog with one end-to-end workflow: fetch PR context, run the swarm, parse/verify the output format, then post comments — including an explicit validation step before any destructive action (request-changes/approve).

Split the agent check-lists, comment templates, GitHub Actions workflow, and config schema into actual files under references/ (e.g. references/agents.md, references/ci-integration.md) and link to them from a concise SKILL.md overview; remove or create the dangling swarm-pr.md and workflow-automation.md links.

Delete the stray duplicate YAML metadata block at the top of the body, fix the corrupted path separators ($ instead of /), and deduplicate the repeated agent/advanced-features sections.

Dimension	Reasoning	Score
Conciseness	The ~540-line body is noticeably padded: the "Review Agents" section restates the "Specialized Review Agents" section as JSON check-lists Claude could infer, "Advanced Review Features" and "Advanced Features" duplicate each other, and placeholder templates ("[link]", "[Clear explanation...]") add tokens without executable value.	2 / 5
Actionability	There are concrete gh CLI commands, but many examples are not executable as written: paths have corrupted separators ("actions$checkout@v3", "$repos/:owner/:repo$pulls/123$comments"), the 19 ruv-swarm invocations rely on undocumented flag combinations with no shown output, and PR numbers are hard-coded without a parameterized pattern.	3 / 5
Workflow Clarity	A rough sequence exists (fetch PR context, initialize swarm, run agents, post results, conditionally approve/request-changes), but there are no validation checkpoints for the batch/destructive PR operations (auto-approving, requesting changes, posting inline comments), which caps this dimension at 3.	3 / 5
Progressive Disclosure	The skill is a monolithic catalog: no references/, scripts/, or assets/ directories exist, the closing "See also" links point to missing files (./swarm-pr.md, ./workflow-automation.md), and content that clearly belongs in separate reference files (comment templates, agent configs, CI/CD workflows, config schema) is fully inlined.	2 / 5
	Total	10 / 20 Passed

Description

28%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 a placeholder-style label: it names the skill but says nothing about what it does, what tasks it covers, or when to invoke it. A rewrite stating the concrete capabilities (multi-agent PR review, security/performance/architecture analysis) plus an explicit 'Use when...' clause would lift all dimensions.

Suggestions

State concrete capabilities, e.g. "Runs multi-agent code reviews on GitHub PRs covering security, performance, style, and architecture findings; posts inline review comments and can request changes on critical issues."

Add an explicit trigger clause: "Use when the user asks to review a pull request, wants an automated or in-depth code review, or mentions PR review, security analysis, or code review agents."

Remove the self-referential invocation text ("invoke with $agent-code-review-swarm") and any meta-labeling that consumes the description budget without describing behavior.

Dimension	Reasoning	Score
Specificity	"Agent skill for code-review-swarm - invoke with $agent-code-review-swarm" names the domain but describes no concrete action or capability; it is a self-referential label rather than a statement of what the skill does.	2 / 5
Completeness	It offers only a vague 'what' ("Agent skill for code-review-swarm") and has no 'when to use' clause at all; the guideline capping a missing 'Use when...' at 3 is not even reached because the 'what' itself is unclear.	2 / 5
Trigger Term Quality	The only keyword is the product name "code-review-swarm"; the natural phrases a user would actually say ("review my PR", "code review", "review this pull request") are missing, matching the one-or-two-generic-keywords anchor.	2 / 5
Distinctiveness Conflict Risk	The tool-specific name gives it a niche, but the generic "Agent skill for..." framing would overlap with any other code-review or agent-orchestration skill and gives no distinct trigger basis.	3 / 5
	Total	9 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (543 lines); consider splitting into references/ and linking	Warning
relative_links	Relative link issues: 4 missing	Warning

	Total	14 / 16 Passed

Repository: ruvnet/ruflo
Path: .agents/skills/agent-code-review-swarm/SKILL.md
Commit: b14c79e

Reviewed: 3 days 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.