agent-coder

Agent skill for coder - invoke with $agent-coder

1.12x

Quality

—

Does it follow best practices?

Impact

79%

1.12x

Average score across 3 eval scenarios

Securityby

Passed

No known issues

Quality

Content

60%

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

The body is a broad 'senior engineer' persona prompt with useful code patterns but heavy generic advice Claude already knows, malformed code tokens, and no progressive disclosure or bundle structure. It reads more like a system prompt than a focused, actionable skill. Tightening the guidance and fixing the broken code would materially raise quality.

Suggestions

Cut generic principles Claude already knows (SOLID/DRY/KISS/YAGNI, coverage thresholds, function-length rules) and keep only skill-specific conventions and examples.

Fix the broken tokens throughout (e.g. '$heavy-module' -> './heavy-module', '$auth$login' -> '/auth/login', 'swarm$coder$status' -> 'swarm/coder/status') so code is copy-paste executable.

Add an explicit validate-then-fix loop to the Implementation Process (e.g. 'Run lint/tests; if they fail, fix and re-run before handing off') and move large reference material (MCP tool examples, file-organization conventions) into separate files referenced one level deep.

Dimension	Reasoning	Score
Conciseness	The body is padded with concepts Claude already knows (SOLID/DRY/KISS/YAGNI, 'keep functions <20 lines', '>80% coverage'), but it also contains concrete code patterns, so it is mostly efficient yet includes unnecessary explanation rather than the lean anchor.	2 / 3
Actionability	It provides real TypeScript code snippets, but several are non-executable or malformed — e.g. 'import(".$heavy-module")', 'api_endpoints: ["$auth$login", "$auth$logout"]', and 'swarm$coder$status' (slash paths rendered as '$') — so guidance is concrete but incomplete/broken.	2 / 3
Workflow Clarity	The 'Implementation Process' lists sequenced steps (Understand, Design, TDD, Incremental), but there are no explicit validation checkpoints or feedback loops despite operations that warrant them (the 'post' hook runs lint, but the workflow body has no validate-then-fix loop).	2 / 3
Progressive Disclosure	It is a monolithic single-file wall of text (~270 lines) with no references, no bundle files in references/scripts/assets, and no navigation to deeper material, matching the monolithic anchor.	1 / 3
	Total	7 / 12 Passed

Description

17%

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 meta text naming the skill category and its invocation token rather than describing capabilities or trigger conditions. It fails to answer what the skill does or when to use it. A complete rewrite with concrete actions and a 'Use when...' clause is needed.

Suggestions

Replace the description with concrete capabilities, e.g. 'Writes, refactors, and optimizes code; designs APIs and implements error handling.'

Add an explicit trigger clause: 'Use when the user asks to write, review, refactor, or optimize code, or to design an API.'

Drop the invocation-token instruction ('invoke with $agent-coder') from the description; it is not a capability and adds no trigger value.

Dimension	Reasoning	Score
Specificity	The phrase 'Agent skill for coder' names no concrete actions — it only labels the category ('coder') and an invocation token, matching the 'vague or no actions' anchor.	1 / 3
Completeness	It says nothing about what the skill concretely does, and there is no 'Use when...' clause, so both 'what' and 'when' are missing or extremely weak.	1 / 3
Trigger Term Quality	The only keyword is 'coder' plus the literal invocation '$agent-coder'; there are no natural trigger terms a user would say, and the text is meta-instruction about invocation rather than user-facing triggers.	1 / 3
Distinctiveness Conflict Risk	'coder' is a broad niche that could overlap with many code-related skills, but the explicit invocation token '$agent-coder' gives it a faint distinguishing handle, placing it just above the fully generic anchor.	2 / 3
	Total	5 / 12 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: ruvnet/claude-flow
Commit: 03d4f84

Reviewed: 2 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.