Content
72%Weight 40%Scale 1-3Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
An actionable, well-structured skill with executable code for all three execution paths and a verification step in the main branch. It loses points for duplicated near-identical scripts and missing validation/feedback loops on the config-write branches.
Suggestions
Collapse the --global and --no-global branches into a single parameterized script (e.g. set global_search to a passed-in boolean) to remove the duplicated python3 block and tighten token use.
Add a feedback loop to the project-name branch: after the search_memories verification, state what to do if zero memories are found or the call errors (e.g. confirm the project_id is correct or guide re-mapping).
Verify the flag branches by reading settings.json back after writing (or printing the resolved global_search value) so the user can confirm the change took effect before restarting.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient with no concept over-explanation, but the --global and --no-global branches ship near-identical python3 blocks that differ only by True/False, and the intro restates the frontmatter description. | 2 / 3 |
Actionability | Every branch provides fully executable, copy-paste-ready python3 -c scripts with exact print outputs, plus a concrete search_memores call with explicit filters and top_k. | 3 / 3 |
Workflow Clarity | Steps are clearly numbered and the project-name branch includes a verification checkpoint, but there is no error-recovery feedback loop and the flag branches lack any verification of the written file. | 2 / 3 |
Progressive Disclosure | Self-contained and cleanly organized into Usage and Execution sub-branches with no nested or dangling references, which is appropriate for a skill of this scope and size. | 3 / 3 |
Total | 10 / 12 Passed |