Content
92%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.
The body is an exemplary lean dispatcher: a single exact command with explicit placeholder resolution and complete, concrete handling of every outcome (success, HALT, missing setup, uv unavailable). Its only minor limitation is that the command is not literally copy-paste ready due to the two placeholders, though their resolution rules are deterministic. Progressive disclosure is handled well, with details correctly split into the bundle and surfaced by the render script.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The 13-line body is lean and assumes Claude's competence: "Run the following command exactly once without changing the current working directory", two one-line placeholder rules, one command, and three bullet rules for outcomes. Every token earns its place with zero padding or explanation of concepts Claude already knows. It is not a 4 because there is nothing that could be trimmed without losing information. | 5 / 5 |
Actionability | The guidance is concrete and near-fully executable — an exact command with explicit, deterministic substitution rules ("Replace `{project-root}` with the nearest folder containing `_bmad/`, from the working directory upward") and complete failure handling for every outcome branch. It is not a 5 because the command is not literally copy-paste ready: the two placeholders require resolution before execution, which constitutes a minor gap; it is well above a 3 because there is no missing step or pseudocode. | 4 / 5 |
Workflow Clarity | For this single-action skill the action is unambiguous (run one command, exactly once, without changing directory) and every outcome has an explicit checkpoint: "`HALT: <reason>`: report the reason and stop", "If the script does not exist... Offer to set it up using `bmad` skill, then run the command again", and "On any other output or failure, including `uv` being unavailable, report the command output and stop". It is not a 4 because the validation/stop conditions and recovery loop are all explicitly present. | 5 / 5 |
Progressive Disclosure | The body is a lean dispatcher while the detail is appropriately split into a real, well-organized bundle — 8 workflow references, 4 asset templates, and 2 scripts (verified to exist, references pointing only one level to scripts, no nested doc chains) — surfaced on demand via "The command prints the skill's workflow: follow it." It is not a 4 because navigation is unambiguous and the split is clean: no inline content that belongs in separate files, no buried or multi-level references. | 5 / 5 |
Total | 19 / 20 Passed |