Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured, concise session wrap-up ritual that clearly sequences its steps and includes appropriate guardrails. The main weakness is the lack of concrete output examples—particularly for the learning capture format and the session summary—which would make the skill more immediately actionable. Overall it's a strong skill that efficiently communicates its purpose.
Suggestions
Add a concrete example of a `[LEARN]` entry, e.g., `[LEARN] Testing: Always run tests with --changed flag to avoid slow full-suite runs`.
Include a brief example of the expected one-paragraph session summary and next-session resume context to make the output format unambiguous.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Every section is lean and purposeful. No unnecessary explanations of what git, linting, or testing are. The learning categories are a compact list, guardrails are terse, and the output spec is a clean bullet list. | 3 / 3 |
Actionability | The bash commands are concrete and executable, but the core workflow steps (Changes Audit, Learning Capture, Next Session Context) are described at a high level without concrete examples of expected output format. For instance, the `[LEARN] Category: Rule` format is mentioned but no example is given. | 2 / 3 |
Workflow Clarity | The five-step workflow is clearly sequenced and logically ordered. Guardrails provide explicit validation checkpoints: flag failing tests before ending, ask about uncommitted changes. The final confirmation prompt ('Ready to end session?') serves as a completion gate. | 3 / 3 |
Progressive Disclosure | This is a simple, single-purpose skill under 50 lines with no need for external references. The content is well-organized into clearly labeled sections (Trigger, Workflow, Commands, Guardrails, Output) that are easy to scan and navigate. | 3 / 3 |
Total | 11 / 12 Passed |