Content
77%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A highly actionable, well-sequenced planning skill with excellent validation checkpoints and a complete output template. Its weaknesses are verbosity in the question-format section and a monolithic structure with no progressive disclosure: the universal QA-format spec and the output template are prime candidates for separate reference files.
Suggestions
Condense the 'Each option stands on its own' bullet in the Question Format section to a compact rule list (name things in words, expand every group pointer, explain coined terms, test by reading the option alone) — the current ~400-word passage repeats the same principle several times.
Move the full 'Output Format' template into a references/ file (e.g., references/epic-template.md) and keep a short summary inline, so the template can be loaded only when writing the file.
Extract the universal 'FR-UNIVERSAL.QA-FORMAT' question-format spec into a shared reference file and link to it, since it is labeled as applying to every question across skills, not just this one.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The rules and step-by-step sections are dense and operational, but the 'Question Format' section — particularly the 'Each option stands on its own' bullet — is a ~400-word passage that restates the same principle (name things in words, not ids or group pointers) many times over. The body is mostly efficient but could be tightened considerably; it does not explain concepts Claude already knows, which keeps it above 2. | 3 / 5 |
Actionability | Guidance is fully concrete and executable: exact frontmatter keys ('date', 'status: to do', 'tags', 'related_tasks'), the exact insertion line '- **Tasks:** [epic-<name>](tasks/...) ', specific commands ('date +%Y-%m-%d', 'mkdir -p'), a complete copy-paste output template, and an idempotency rule for the SRS edit. For an instruction-only skill this covers the common cases completely. | 5 / 5 |
Workflow Clarity | The 8 steps are clearly sequenced with explicit checkpoints ('STOP and wait' in steps 2 and 4), a critique-then-refine feedback loop (steps 6-7), byte-identical surgical-edit constraints, and a closing verification checklist. This matches the top anchor: clear sequence, explicit validation, feedback loops, and a checklist. | 5 / 5 |
Progressive Disclosure | The skill is a ~220-line monolith with no bundle files at all; content that plausibly belongs in separate references — the universal 'FR-UNIVERSAL.QA-FORMAT' question-format spec and the full output template — is inlined. Section structure itself is good (When to Use / Context / Rules / Instructions / Output Format / Verification), and it leans on an external AGENTS.md rather than bundled references, which keeps it above 2. | 3 / 5 |
Total | 16 / 20 Passed |