Content
39%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill demonstrates excellent workflow design with clear multi-phase sequencing, conditional branching, and validation gates, but suffers severely from verbosity and monolithic structure. The ~400+ lines of inline JavaScript implementation details consume excessive context window budget and should be split into referenced files. Several code sections are incomplete or contain errors, reducing actionability despite the generally concrete approach.
Suggestions
Extract phase implementations into separate files (e.g., phases/phase-1-session.md, phases/phase-4-task-gen.md) and keep SKILL.md as a concise overview with the pipeline diagram and references
Fix incomplete code: define the `conflicts` variable in Phase 3, fix the syntax error in Phase 5's if condition, and replace pseudocode comments like '// Wait for CLI → conflicts[]' with actual implementation or explicit TODOs
Remove redundant content: the ASCII pipeline diagram and the Data Flow diagram convey overlapping information — keep one and cut the other
Trim session discovery logic significantly — the multi-session selection, slug generation, and folder creation code is implementation detail that doesn't need to be in the skill body
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~400+ lines. It includes extensive inline code that could be in separate files, redundant ASCII diagrams (pipeline + data flow), and over-explains session discovery logic. Much of the JavaScript is implementation detail that bloats the context window without adding proportional instructional value. | 1 / 3 |
Actionability | The code examples are fairly concrete but not fully executable — Phase 3 references undefined variables (conflicts), Phase 5 has syntax errors (bare `||` in if condition), and several sections use pseudocode comments like '// Wait for CLI → conflicts[]' or '// Wait for CLI → verification report' without actual implementation. The spawn_agent/wait_agent API usage is concrete but the CLI integration (ccw cli) sections are incomplete. | 2 / 3 |
Workflow Clarity | The multi-phase pipeline is exceptionally well-sequenced with clear phase numbering, conditional branching (conflict risk gating Phase 3), explicit validation checkpoints (Plan Confirmation Gate, Phase 5 verification with 10 dimensions), backup before replan, and an error handling table. The feedback loop for verify→replan is clearly articulated. | 3 / 3 |
Progressive Disclosure | This is a monolithic wall of text with no bundle files and no references to external documents. All implementation details, data flow diagrams, session structures, and phase implementations are inlined into a single massive file. The session structure diagram, full JavaScript implementations for all 6 phases, and the ASCII pipeline diagram should be split into separate referenced files. | 1 / 3 |
Total | 7 / 12 Passed |