Content
85%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 well-structured, highly actionable dispatch skill with clear sequencing, explicit validation and repair feedback loops, and clean one-level-deep reference navigation. The only soft spots are a few illustrative/placeholder commands and a maintainer-only notes section adding non-execution tokens.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense and mostly token-efficient — tables, terse operational prose, and concrete commands with little concept-explanation fluff — but the maintainer-only 'Design notes' section and a few routing/intent prose passages could be trimmed. | 4 / 5 |
Actionability | Concrete, copy-paste-ready commands for every Bash phase ('npx hyperframes init', 'lint', 'check', 'snapshot', 'preview', 'render') plus a validation one-liner and a JSON schema; minor gaps remain (the source command is explicitly 'illustrative' and several commands carry placeholders like '<proof-times>', '<SKILL_DIR>'). | 4 / 5 |
Workflow Clarity | Steps 0–6 are clearly sequenced with explicit validation checkpoints (Step 1 existence check, Step 5 lint/check/snapshot with a repair feedback loop that re-runs the failed gate) and a resume table for state recovery — satisfying the destructive/batch feedback-loop expectation. | 5 / 5 |
Progressive Disclosure | The body is a dispatch overview pointing one level deep to clearly signaled detail files (agents/*.md, phases/source/guide.md, categories/<id>/module.md, references/motion-vocabulary.md, shot-plan-ir.md, builder-contract.md, catalog-map.md), and the bundled references that exist are real and referenced. | 5 / 5 |
Total | 18 / 20 Passed |