Content
7%Scale 1-3Reviews 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, presenting a sprawling collection of hypothetical CLI commands and aspirational features rather than actionable, executable guidance. Most code examples reference a `ruv-swarm` tool with undocumented commands, making them non-executable. The content lacks clear workflows with validation checkpoints, and the massive redundancy between sections (especially the 'Advanced Swarm PR Coordination' section repeating earlier concepts) wastes significant token budget.
Suggestions
Reduce content to under 100 lines by eliminating redundant sections and focusing on the actual executable workflow: identify which commands/tools actually exist and document only those with working examples.
Replace hypothetical `npx ruv-swarm` commands with actual MCP tool calls using correct syntax, or clearly document the tool's installation and API if it's a real dependency.
Add a clear sequential workflow with explicit validation checkpoints, e.g.: 1. Analyze PR diff → 2. Spawn review agents → 3. Validate results → 4. Post review → 5. Merge decision, with error handling at each step.
Fix syntax errors throughout: GitHub Actions paths use `$` instead of `/`, MCP tool calls use invalid JSON-like syntax, and the webhook handler is missing request body parsing.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at ~350+ lines with massive redundancy. Multiple sections show the same concepts (PR init, review, merge) repeated with slight variations. The 'Advanced Swarm PR Coordination' section largely duplicates earlier content using MCP tool syntax. Explains obvious concepts like webhook handling and label mapping that Claude already understands. Many code blocks are aspirational/non-executable CLI commands for a tool (ruv-swarm) without clear documentation of its actual API. | 1 / 3 |
Actionability | Most code examples reference `npx ruv-swarm` commands that appear to be hypothetical/aspirational rather than executable. The MCP tool invocations use incorrect syntax (JSON-like objects instead of actual tool call format). The webhook handler is incomplete (missing body parsing). The GitHub Actions workflow uses `$` instead of `/` in paths. Very little is copy-paste ready or verifiably executable. | 1 / 3 |
Workflow Clarity | No clear sequential workflow with validation checkpoints. The skill presents a grab-bag of features and commands without a coherent process flow. The 'Integration with Claude Code' section lists 5 steps but they're vague and lack validation. For destructive operations like auto-merge, there are no explicit validation/rollback steps. The pre/post hooks in frontmatter describe intentions but the body doesn't implement them clearly. | 1 / 3 |
Progressive Disclosure | References to related files (swarm-issue.md, sync-coordinator.md, workflow-automation.md) at the bottom provide some navigation, but no bundle files exist to support them. The content itself is monolithic with everything inline rather than appropriately split. Some section organization exists (Core Features, Advanced Features, Best Practices) which prevents a score of 1. | 2 / 3 |
Total | 5 / 12 Passed |