Content
92%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 tight, highly actionable skill body with correct edge-case handling and zero wasted tokens. Its only defects are a self-contradiction between the Step 3 markdown-header example and the no-markdown output rule, and undefined placeholders for the required user_id/app_id filter values.
Suggestions
Make the Step 3 display template comply with the no-markdown rule: replace the '## mem0 search: "<query>" (<N> results)' header line with a plain-text equivalent so the example doesn't contradict the Output formatting section.
Define where the user_id and app_id filter values come from (e.g., project context or a stated convention) so the search_memories call is fully executable without guessing.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Lean and efficient: every section carries non-obvious information (exact regexes, exact call parameters, output template, the environment-specific TUI rendering rule). Nothing explains concepts Claude already knows; no padding. The example block is the minimal illustration needed to pin the format. | 5 / 5 |
Actionability | Mostly fully executable: exact regex patterns, a concrete search_memories call with filters and top_k, a precise output format spec, and an example. Two minor gaps keep it below 5: the filter placeholders "<id>" and "<pid>" are never defined (where do user_id/app_id values come from?), and the Step 3 display template opens with a markdown '## mem0 search:' header that contradicts the 'Do NOT use markdown' instruction. | 4 / 5 |
Workflow Clarity | The sequence (parse query → ID lookup with explicit not-found fallthrough → search → display) is unambiguous, with all edge cases handled: no query provided, bare hex/UUID/citation input, ID miss, and zero results. This is a read-only single-purpose skill, so no destructive/batch validation cap applies, and the fallthrough path is a genuine error-recovery loop. | 5 / 5 |
Progressive Disclosure | Under 50 lines, single task, no bundle files and no content that warrants offloading — well-organized into Execution steps and Output formatting sections. Per the simple-skill guideline, this scores 5 with just well-organized sections. | 5 / 5 |
Total | 19 / 20 Passed |