Content
100%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.
An exemplary process skill body: dense but unambiguous, fully actionable commands, explicit sequencing with validation checkpoints and fallbacks, and clean separation of overview from one-level-deep reference material. No dimension shows a meaningful weakness.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is relentlessly lean — e.g. "Sections omittable; order locked when present." — with no explanation of concepts Claude already knows and no padding; even illustrations like the named-protagonist examples exist to pin a rule ("Mary, mom of three, kids asleep — not 'the user'"), so every token earns its place. | 5 / 5 |
Actionability | Guidance is fully concrete and executable: exact commands with flags ("uv run {project-root}/_bmad/scripts/memlog.py append --workspace {doc_workspace} --type <decision|change|override|assumption|event> --text \"…\""), exact folder patterns ("ux-{slug}/ux-{slug}.md"), and specific fallback steps for each failure mode (script not found, no active_initiative). | 5 / 5 |
Workflow Clarity | Multi-step processes are explicitly sequenced (numbered On Activation 1–5, ordered Finalize outcomes) with real checkpoints — the Reviewer Gate opt-in, "Runs the rubric walker's Pass 1 coverage checks proactively", the mock-coverage walk, and "confirm every entry was executed in order" — plus error-recovery paths (resume detection, misroute scan, script-not-found fallbacks). | 5 / 5 |
Progressive Disclosure | The body is an overview that pushes all detail to real, clearly-signaled, one-level-deep bundle files — references/design-md-spec.md, references/validate.md, references/headless.md, and references/creative-tools.md all exist, with examples and templates in assets/ — and references are cited exactly where each is needed. | 5 / 5 |
Total | 20 / 20 Passed |