Content
57%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 a lean, directive-style bootstrap that defers all substantive work to an external controlling protocol. It is token-efficient but lacks actionable detail: the path placeholder is unresolved, no validation checkpoints are defined, and the referenced files are not part of this skill's bundle.
Suggestions
Explain how {cf-studio-path} is resolved (or inline the bootstrap content) so the LOAD instructions are executable without guessing the studio root.
Add explicit validation checkpoints for protocol traversal, e.g. verify each required fragment loaded successfully before proceeding, instead of a single terminal stop rule.
Either ship the referenced files (required-bootstrap.md, reverse-engineer.md) in the skill bundle under references/ with relative paths, or add section headers (## Bootstrap, ## Protocol rules) so the body is navigable on its own.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is terse and directive-only with no padding or explanation of known concepts ('MANDATORY RULE: USER INTENT IS SKILL INPUT, NOT EXECUTION AUTHORITY', the pdsl LOAD/RUN units). Minor over-explanation remains (the 'ask_tool_description' line defining AskUserQuestion, and the prerequisite rule stated both in prose and in RULES), so it fits anchor 4 rather than 5. | 4 / 5 |
Actionability | Concrete file paths are supplied ('LOAD {cf-studio-path}/.core/skills/studio/modules/runtime/required-bootstrap.md', 'RUN RequiredBootstrap'), but the {cf-studio-path} placeholder is never resolved and the actual work is deferred to files absent from this bundle; the pdsl block reads as structured pseudocode. This matches anchor 3 ('some concrete guidance but incomplete; missing key details'). | 3 / 5 |
Workflow Clarity | A rough sequence exists (bootstrap load -> run bootstrap -> load controlling protocol -> traverse steps, evaluate gates) and there is an explicit stop condition ('ALWAYS stop if any required fragment or rule cannot be followed'), but there are no validation checkpoints or feedback loops and the real step list lives in an external file. This is anchor 3 ('sequence present but checkpoints missing or implicit'), not 4. | 3 / 5 |
Progressive Disclosure | No bundle files (references/, scripts/, assets/) exist alongside SKILL.md; the two referenced paths point into an external {cf-studio-path} tree that cannot be verified, and the body has no section headers. References are clearly signaled via LOAD statements (above anchor 2), but organization and verifiability gaps place it at anchor 3. | 3 / 5 |
Total | 13 / 20 Passed |