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 instructional skill: a clearly sequenced 9-stage pipeline with explicit validation checkpoints and feedback loops, supported by well-signaled one-level-deep references. It is mostly lean and actionable, with only minor trimming and a few inline-template gaps keeping conciseness and actionability at 4.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean and assumes Claude's competence (e.g., 'a knowledge graph is a product with a schema, not a pile of triples'), but the Credits section and some repetition between inline stage descriptions and the Reference Files list could be trimmed; not a 5 because not every token earns its place. | 4 / 5 |
Actionability | Concrete, specific directives throughout ('5-15 entity types, 10-30 relation types', 'every relation gets a precise verb name (ACQUIRED, not RELATED_TO)', 'Target ≥90% precision on a 50-item sample') make it actionable; not a 5 because executable prompt templates/code are deferred to references rather than inline, leaving minor gaps. | 4 / 5 |
Workflow Clarity | The 9-stage pipeline is explicitly sequenced with a dedicated quality-gate checkpoint (stage 7: sample, score ≥90% precision, 'fix the prompt/rules not the output, then re-run') plus a '10-document pilot before scaling' gate; matches the anchor-5 example of clear sequence with explicit validation and feedback loops. | 5 / 5 |
Progressive Disclosure | Clear overview with five one-level-deep reference files, each clearly signaled with 'Read when...' guidance and stage associations, all real files present in ./references/; content appropriately split and easy to navigate, matching the anchor-5 example. | 5 / 5 |
Total | 18 / 20 Passed |