Content
0%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is an extremely verbose, largely aspirational document that reads more like a product marketing page than actionable instructions. The code examples mix real shell commands with invented DSL/pseudocode that wouldn't execute, and destructive batch operations across entire organizations lack any meaningful validation or rollback mechanisms. The 500+ line monolithic structure with no supporting bundle files makes it impractical as a skill reference.
Suggestions
Reduce to under 100 lines by keeping only the 2-3 most critical workflows with truly executable code, and move everything else to referenced bundle files
Replace pseudocode and hypothetical API calls (Task(), mcp__claude-flow__*, etc.) with actual executable commands or clearly mark them as requiring specific tool installations with verification steps
Add explicit validation checkpoints and rollback procedures for all destructive operations (org-wide updates, security patches, cross-repo pushes) - e.g., dry-run mode, diff review before commit, rollback scripts
Split into a concise SKILL.md overview with references to separate files for configuration (config.md), workflows (workflows.md), architecture patterns (architecture.md), and troubleshooting (troubleshooting.md)
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at 500+ lines with massive amounts of redundant content. Many sections are aspirational CLI commands that appear to be hypothetical rather than functional. Extensive repetition of patterns (swarm init, task spawning, etc.) and sections like 'Best Practices', 'Metrics and Reporting', 'Support and Resources' add little actionable value. Explains concepts Claude already knows and includes decorative emoji headers. | 1 / 3 |
Actionability | Despite containing many code blocks, the code is largely pseudocode or hypothetical API calls (e.g., `mcp__claude-flow__swarm_init`, `Task()`, `Read()`) that are not standard executable code. The CLI commands like `npx claude-flow skill run github-multi-repo microservices` appear to be aspirational rather than real commands. The JavaScript blocks mix real shell commands with invented DSL syntax that wouldn't actually run. | 1 / 3 |
Workflow Clarity | Multi-step processes lack explicit validation checkpoints and error recovery. The synchronized operations section runs destructive batch operations (cloning repos, pushing changes, creating PRs) with only a basic `$?` check. No feedback loops for error recovery, no rollback procedures, and no verification that changes are correct before pushing. For operations this destructive (org-wide dependency updates, security patches), this is critically insufficient. | 1 / 3 |
Progressive Disclosure | Monolithic wall of text with no bundle files to reference. Everything is inlined into a single massive document. Content that should be in separate files (configuration examples, architecture patterns, troubleshooting guides, individual workflow details) is all crammed into one file. No external references are provided despite the document being far too long for a single skill file. | 1 / 3 |
Total | 4 / 12 Passed |