Content
62%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is exceptionally thorough and actionable with excellent workflow clarity, explicit error handling, and concrete instructions at every step. However, it is severely over-long — the inline detail for every sub-flow, edge case, and preference lookup makes it a massive token consumer that would benefit greatly from extracting major subsections (scope assessment, complex phase upgrade gate, swift/standard plan generation) into reference files. The content is high quality but poorly budgeted for context window efficiency.
Suggestions
Extract the Complex Phase Upgrade Gate (Step 3.5c, ~60 lines) into a reference file like `references/complex-phase-upgrade-gate.md` and replace with a 2-line reference pointer
Move the scope assessment scoring procedure (Step 2.5b, detailed criterion weighting and override rules) into a reference file, keeping only the high-level flow inline
Extract the swift/standard plan generation logic (Step 2.6) into a reference file — the template loading, directory creation, and review flow details don't need to be in the main sequencer
Consolidate the repeated AskUserQuestion patterns — many follow identical structures (3 options: proceed/switch/exit) that could be referenced as a standard pattern rather than spelled out each time
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | This skill is extremely verbose at ~400+ lines with extensive inline logic for every edge case, sub-option, and error path. Much of this could be delegated to reference files or condensed significantly. The multi-level selection flows, complex phase upgrade gate with session memory model, and tier override logic are all spelled out in exhaustive detail that bloats the token budget enormously. | 1 / 3 |
Actionability | The skill provides highly specific, concrete instructions at every step: exact commands (gh issue list), exact file paths, exact AskUserQuestion option text, exact table formats, and precise branching logic. Every decision point has explicit options and outcomes, making it fully executable. | 3 / 3 |
Workflow Clarity | The multi-step workflow is clearly sequenced (Steps 0-6) with explicit validation checkpoints, error handling with retry/skip/abort options at each failure point, status reversion on failure, progress reporting, and feedback loops throughout. The pipeline diagram at the top provides excellent orientation. | 3 / 3 |
Progressive Disclosure | The skill does reference external files appropriately (backlog-selection.md, plan-shipping.md, scope-router-criteria.md, templates), but an enormous amount of procedural logic is inline that could be in reference files — particularly the Complex Phase Upgrade Gate (Step 3.5c), the scope assessment scoring procedure (Step 2.5), and the swift/standard plan generation (Step 2.6). The main file tries to be both overview and complete procedure. | 2 / 3 |
Total | 9 / 12 Passed |