Unified memory capture with routing - session compact or quick tips. Triggers on "memory capture", "compact session", "save session", "quick tip", "memory tips", "记录", "压缩会话".
49
53%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Risky
Do not use without reviewing
Optimize this skill with Tessl
npx tessl skill review --optimize ./.claude/skills/memory-capture/SKILL.mdQuality
Discovery
72%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
The description provides good trigger term coverage including multilingual keywords, and occupies a distinct niche. However, it lacks specificity about what concrete actions the skill performs (e.g., what does 'routing' mean, what does a compacted session look like, what are 'quick tips'?). Adding a 'Use when...' clause with situational context and more concrete action descriptions would strengthen it.
Suggestions
Add concrete action descriptions — e.g., 'Compresses session conversation into structured memory notes' or 'Saves reusable tips extracted from the current session' instead of just 'session compact or quick tips'.
Include a 'Use when...' clause that describes the situational context, e.g., 'Use when the user wants to preserve key learnings from a session or save concise tips for future reference.'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain ('memory capture with routing') and mentions two specific actions ('session compact' and 'quick tips'), but doesn't elaborate on what these actions concretely entail — e.g., what is being compacted, what format tips are saved in, etc. | 2 / 3 |
Completeness | The 'what' is partially addressed ('memory capture with routing - session compact or quick tips') but remains vague. The 'when' is addressed via explicit trigger terms, but there is no 'Use when...' clause that describes the situational context — only keyword triggers are listed. This is borderline but the explicit trigger list partially compensates. | 2 / 3 |
Trigger Term Quality | Includes a good set of natural trigger terms users would say: 'compact session', 'save session', 'quick tip', 'memory tips', 'memory capture', plus multilingual support with '记录' and '压缩会话'. These cover common variations well. | 3 / 3 |
Distinctiveness Conflict Risk | The combination of 'memory capture', 'session compact', and 'quick tips' with routing logic creates a clear niche. The specific trigger terms are distinctive enough to avoid conflicts with general note-taking or session management skills. | 3 / 3 |
Total | 10 / 12 Passed |
Implementation
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill has a clear routing concept and good error handling documentation, but suffers significantly from verbosity — the same routing logic is expressed in at least four different ways (ASCII diagram, routing table, data flow diagram, prose). The actual execution details are deferred to phase files that aren't provided, making the skill more of an architectural description than actionable guidance. Trimming redundant representations and ensuring phase files exist would substantially improve quality.
Suggestions
Remove redundant representations: pick ONE of the ASCII architecture diagram, the data flow diagram, or the auto-route table — not all three. The routing table alone is sufficient.
Provide the referenced phase files (phases/01-compact.md, phases/02-tips.md) or inline their key content, since without them the skill is not self-sufficient for execution.
Add a validation step after structured text generation and before core_memory import to verify the output meets expected format/length constraints.
Remove the 'Architecture Overview' ASCII diagram entirely — it adds ~15 lines that duplicate information already conveyed by the routing table and step descriptions.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is verbose with multiple redundant representations of the same routing logic (ASCII diagram, auto-route table, data flow diagram, and prose). The architecture overview diagram, execution flow, and data flow sections all convey essentially the same information. Claude doesn't need concepts like 'single phase execution' or 'data flow' explained with diagrams. | 1 / 3 |
Actionability | The core_memory import call is concrete and executable, and the routing rules table is specific. However, the actual execution logic is deferred to phase files (01-compact.md, 02-tips.md) which are not provided, making the skill incomplete on its own. The input processing examples are helpful but the skill mostly describes architecture rather than giving executable instructions. | 2 / 3 |
Workflow Clarity | The three-step workflow (parse/route → execute phase → save) is clearly sequenced, and error handling is documented in a table. However, there are no validation checkpoints between steps — no verification that the structured text was correctly generated before importing, and the retry logic for core_memory failure is mentioned but not detailed with a feedback loop. | 2 / 3 |
Progressive Disclosure | The skill correctly references phase files (phases/01-compact.md, phases/02-tips.md) for detailed execution, which is good progressive disclosure. However, no bundle files are provided, so we can't verify these references exist. The main file itself contains too much redundant content (architecture diagram + data flow + execution flow all covering the same routing) that should either be trimmed or moved to a reference file. | 2 / 3 |
Total | 7 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
allowed_tools_field | 'allowed-tools' contains unusual tool name(s) | Warning |
Total | 10 / 11 Passed | |
5ff5e86
Table of Contents
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.