agent-swarm-pr
Agent skill for swarm-pr - invoke with $agent-swarm-pr
41
10%
Does it follow best practices?
Impact
97%
2.62xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./.agents/skills/agent-swarm-pr/SKILL.mdQuality
Discovery
0%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an extremely weak description that fails on every dimension. It provides no information about what the skill does, when it should be used, or what user requests should trigger it. It reads more like a stub or placeholder than a functional skill description.
Suggestions
Add concrete actions describing what swarm-pr actually does (e.g., 'Creates and manages pull requests across multiple repositories', 'Coordinates parallel PR reviews').
Add an explicit 'Use when...' clause with natural trigger terms users would say (e.g., 'Use when the user asks about pull requests, PRs, code review, or merging changes').
Remove the invocation instruction ('invoke with $agent-swarm-pr') from the description—this is operational detail, not selection criteria—and replace it with capability and context information.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description contains no concrete actions whatsoever. 'Agent skill for swarm-pr' is entirely vague—it doesn't describe what the skill does, only names itself. There are no verbs describing capabilities. | 1 / 3 |
Completeness | Neither 'what does this do' nor 'when should Claude use it' is answered. The description only states it's an 'agent skill' and how to invoke it, providing no functional or contextual information. | 1 / 3 |
Trigger Term Quality | The only potentially relevant term is 'swarm-pr', which is a technical/internal name rather than a natural keyword a user would say. There are no natural language trigger terms like 'pull request', 'review', 'PR', etc. | 1 / 3 |
Distinctiveness Conflict Risk | The description is so generic ('agent skill') that it provides almost no distinguishing information. Without knowing what it does, it could conflict with any number of other skills or simply never be selected appropriately. | 1 / 3 |
Total | 4 / 12 Passed |
Implementation
20%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is excessively verbose and largely non-actionable, consisting primarily of hypothetical CLI commands for a tool ('npx ruv-swarm') that appears aspirational rather than real. The MCP tool call syntax is incorrect, code examples are not executable, and the same concepts (PR initialization, agent spawning) are repeated across multiple sections without adding value. The content would benefit from being reduced by 70%+ and focused on actual executable workflows using the declared MCP tools.
Suggestions
Remove all 'npx ruv-swarm' commands and replace with actual executable workflows using the declared MCP tools (mcp__github__*, mcp__claude-flow__*) with correct invocation syntax
Reduce content by at least 70% — consolidate the 5+ PR initialization examples into one canonical workflow with clear steps and validation checkpoints
Add explicit validation and error recovery steps (e.g., what to do when PR review fails, when merge conflicts arise, when agent spawning fails) to create proper feedback loops
Move GitHub Actions YAML, webhook handler, and label mapping configuration into separate referenced files to improve progressive disclosure
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at ~300+ lines with massive redundancy. Multiple sections repeat the same concepts (PR initialization shown 5+ times in different forms). Includes explanations Claude doesn't need (what webhooks are, how GitHub Actions work), and many code blocks are aspirational/non-executable examples of a tool ('npx ruv-swarm') that appear to be hypothetical CLI commands rather than real actionable guidance. | 1 / 3 |
Actionability | Most code examples reference a hypothetical CLI tool ('npx ruv-swarm') with commands that appear fabricated rather than executable. The MCP tool invocations use incorrect syntax (JSON-like objects instead of actual tool call format). The webhook handler is incomplete (missing body parsing). Very little is copy-paste ready or verifiably executable. | 1 / 3 |
Workflow Clarity | There is a general sequence implied (init → spawn agents → review → merge), and the Claude Code integration section lists 5 steps. However, there are no explicit validation checkpoints or error recovery steps. The merge automation section has some conditional checks but lacks feedback loops for when things fail. | 2 / 3 |
Progressive Disclosure | References to related files (swarm-issue.md, sync-coordinator.md, workflow-automation.md) exist at the bottom, which is good. However, the main file is a monolithic wall of content that should be split — the webhook handler, GitHub Actions config, label mapping, and advanced coordination sections could all be separate files. Too much inline content that buries the core workflow. | 2 / 3 |
Total | 6 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- ruvnet/claude-flow
- Commit
f547cec
- Reviewed
Table of Contents
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.