Content
70%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a well-structured meta-skill that effectively serves as a routing layer for discovering and applying other skills. Its main strengths are the clear decision tree, lifecycle sequencing, and progressive disclosure to individual skills. Its weaknesses are moderate redundancy across three different representations of the same skill catalog (tree, lifecycle, table) and some verbosity in the core behaviors section that explains concepts Claude already understands.
Suggestions
Consolidate the three overlapping skill representations (discovery tree, lifecycle sequence, quick reference table) — pick two at most, as all three convey largely the same information and consume significant tokens.
Add concrete guidance on how to actually load/invoke a skill (e.g., file paths, exact commands like 'read skills/<skill-name>/SKILL.md') rather than just naming skills abstractly.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is reasonably well-structured but includes some redundancy — the lifecycle sequence, quick reference table, and skill discovery tree all convey overlapping information. The 'Core Operating Behaviors' section, while valuable, is somewhat verbose with explanations of concepts Claude likely already understands (e.g., explaining why sycophancy is bad). The 'Failure Modes to Avoid' section largely restates the core behaviors in negative form. | 2 / 3 |
Actionability | The skill discovery flowchart is concrete and actionable for routing tasks to skills. The core operating behaviors provide good templates (e.g., the ASSUMPTIONS format). However, the skill lacks executable commands or concrete examples of how to actually invoke/load a skill — it describes the process abstractly rather than showing exact steps like 'read SKILL.md from skills/<skill-name>/'. | 2 / 3 |
Workflow Clarity | The decision tree for skill discovery is clear and well-sequenced. The lifecycle sequence provides explicit ordering. The core behaviors include clear validation checkpoints ('verify, don't assume') and feedback loops ('STOP, name confusion, wait for resolution'). The rules about not skipping verification steps are explicit. | 3 / 3 |
Progressive Disclosure | This is a meta-skill that serves as an overview/index pointing to 20+ individual skills, each referenced by name. The structure naturally provides progressive disclosure — quick reference table for scanning, decision tree for routing, and individual skill files for detailed workflows. Content is appropriately split between overview (here) and detail (individual skill files). | 3 / 3 |
Total | 10 / 12 Passed |