CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

memory-config

Diagnose and configure MemSearch memory behavior. Use when the user asks about MemSearch configuration, plugin summarization, PROJECT.md/USER.md maintenance, memory directories, index health, provider routing, prompt files, or migration/compatibility questions.

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Critical

Do not install without reviewing

The canonical home for this skill is memory-config in zilliztech/memsearch

SKILL.md
Quality
Evals
Security

Quality

Content

78%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-engineered configuration skill body: executable diagnostic commands, concrete config semantics, careful safety gating (ask before enabling paid providers, re-indexing, or deleting state), and exemplary progressive disclosure of platform details into real reference files. The main improvements are consolidating repeated disclaimers/platform-file pointers, showing a concrete "memsearch config set" example, and making the diagnose→fix→re-verify loop explicit.

Suggestions

Add one concrete example of the primary change command, e.g., `memsearch config set plugins.claude-code.summarize.provider native` (global) and an allowlisted `--project` variant, since "Applying changes" currently names the command without showing it.

State the repeated "this is MemSearch memory configuration, not the host agent's own memory/config system" disclaimer once (as the opening already does) and drop the duplicate in "Applying changes", and consider consolidating the six "see your platform reference file" pointers into one clearly marked section.

Make the troubleshooting feedback loop explicit for index errors: after fixing the cause of a degraded/error status, re-run `memsearch stats` / re-check `.index-state.json` to confirm the status is healthy before reporting success.

DimensionReasoningScore

Conciseness

The body is dense and almost entirely MemSearch-specific (config allowlists, resolution order, state-file semantics, provider rules) with no explanation of concepts Claude already knows. Minor trimmable repetition — the "not the host agent's own memory/config system" disclaimer appears twice and "see your platform reference file" recurs roughly six times — keeps it at anchor 4 rather than 5.

4 / 5

Actionability

Diagnostics are copy-paste ready ("memsearch config list --resolved", the version-check pipeline, memory-file and index-state inspection blocks) and provider routing includes concrete TOML examples with realistic model names. The gap is that "memsearch config set" — the core change operation — is never demonstrated with a concrete example invocation, so it sits at anchor 4 rather than 5.

4 / 5

Workflow Clarity

A clear sequence is present (inspect request → route intent → read platform file → diagnose first → apply changes → show resolved value and restart needs) with post-change verification and error-state interpretation ("status: degraded means the scan completed but one or more files failed", .maintenance-state.json last_error fields). It falls short of anchor 5 because the sequence must be synthesized across sections rather than followed as one ordered workflow, and there is no explicit re-verify loop after fixing a degraded index.

4 / 5

Progressive Disclosure

Platform-specific detail is correctly split into six real, substantive, one-level-deep reference files (references/workbuddy.md through references/dsh.md, all verified to exist), and the body instructs reading only the file matching the current environment with explicit environment markers for disambiguation. This matches anchor 5: clear overview, well-signaled one-level-deep references, easy navigation.

5 / 5

Total

17

/

20

Passed

Description

78%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description with an explicit third-person what-and-when structure and good natural trigger terms anchored to a named tool. Its main weaknesses are limited action specificity (only "diagnose and configure") and residual conflict risk with the host agent's own memory/config skills via generic memory terms.

Suggestions

Enumerate one or two more concrete capabilities after "Diagnose and configure" (e.g., "...troubleshoot index health, route LLM providers, and override prompt files") to raise specificity.

Add the natural user phrasings that are currently missing, such as "memory search", "recall returns nothing", or "API key/provider setup", to broaden trigger coverage.

Qualify the generic memory terms to reduce conflict with the host agent's built-in memory system, e.g., "MemSearch memory directories (.memsearch/)" instead of bare "memory directories".

DimensionReasoningScore

Specificity

The description names the domain and exactly two concrete actions — "Diagnose and configure MemSearch memory behavior" — matching the anchor for 1-2 concrete actions without comprehensive coverage. It does not list the several specific sub-capabilities the skill covers (index troubleshooting, provider setup, prompt overrides, migration), so it falls short of anchor 4.

3 / 5

Completeness

It clearly answers both questions: the "what" is stated upfront ("Diagnose and configure MemSearch memory behavior") and the "when" is explicit with concrete trigger phrases ("Use when the user asks about MemSearch configuration, plugin summarization, ... migration/compatibility questions"), in third-person voice. This matches anchor 5 exactly.

5 / 5

Trigger Term Quality

Strong natural keyword coverage across sub-topics: "MemSearch configuration", "plugin summarization", "PROJECT.md/USER.md maintenance", "memory directories", "index health", "provider routing", "prompt files", "migration/compatibility questions". A few natural user phrasings are missing (e.g., "memory search", "recall returns nothing", "API key setup"), keeping it below anchor 5's comprehensive synonym coverage.

4 / 5

Distinctiveness Conflict Risk

MemSearch is a named niche tool with distinctive artifacts (PROJECT.md/USER.md, index health), giving a clear niche. However, generic terms like "memory directories" and "memory behavior" risk triggering for questions about the host agent's own memory system — a confusion the skill body itself repeatedly guards against — so minor overlap risk keeps it at anchor 4 rather than 5.

4 / 5

Total

16

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

zilliztech/memsearch

Path

plugins/_shared/skills/memory-config/SKILL.md

Reviewed

Table of Contents

Is this your skill?

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.