CtrlK
BlogDocsLog inGet started
Tessl Logo

minicpm5-deploy-lmstudio

Run MiniCPM5-1B in LM Studio (desktop GUI) using either the GGUF runtime (cross-platform) or the MLX runtime (Apple Silicon, faster). Includes OpenAI-compatible local server. Use when the user mentions "LM Studio", desktop GUI inference, "lms" CLI, or wants a no-code chat UI for MiniCPM5.

72

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

90%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.

The content is highly actionable and token-efficient, with concrete commands for both runtimes and explicit validation. The main gap is the absence of an explicit error-recovery loop after the validation step.

Suggestions

Add a brief feedback loop after Step 3 (Validate): e.g. "If the reply is not '2', confirm the model loaded (`lms ps`) and that the correct runtime path was taken before retrying."

Either inline a minimal version of the referenced docs/deployment/lmstudio.md or confirm the path resolves, since the bundle contains no references/ directory to back the pointer.

DimensionReasoningScore

Conciseness

The body is lean and dense — tables for runtimes and inputs, copy-paste bash blocks, and only non-obvious asides (which MLX repo exists, the LM Studio 0.4.13 think-mode gotcha) — with no padding explaining concepts Claude already knows.

5 / 5

Actionability

Every step is copy-paste executable: brew install, huggingface-cli download with concrete paths, lms server/load/ps commands, and a complete curl validation request with a JSON body, covering both GGUF and MLX paths.

5 / 5

Workflow Clarity

Steps are clearly sequenced (install → 2A/2B runtime path → validate) with explicit validation checkpoints (lms ps, curl with expected "2"), but there is no explicit "if validation fails, fix and re-run" feedback loop, which the score-5 anchor requires.

4 / 5

Progressive Disclosure

Well-organized into clear sections (runtimes, inputs, steps, think/nothink, pitfalls, when-not-to-use) with a single clearly signaled one-level reference (docs/deployment/lmstudio.md), though the referenced doc lives outside the skill bundle and cannot be verified.

4 / 5

Total

18

/

20

Passed

Description

87%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.

The description is well-crafted: it answers both what and when with concrete trigger phrases, names specific runtimes and features, and carves out a distinct niche. Minor gains are possible by adding a couple more synonyms (e.g. "local server").

Suggestions

Add "local server" or "OpenAI-compatible API" to the trigger phrases so users describing server use cases land on this skill.

Consider mentioning the quantization options (Q4_K_M / 4-bit) briefly to differentiate from sibling deployment skills that target other runtimes.

DimensionReasoningScore

Specificity

Names the domain (MiniCPM5-1B in LM Studio) and lists several concrete specifics — GGUF runtime (cross-platform), MLX runtime (Apple Silicon, faster), and an OpenAI-compatible local server — comparable in concreteness to the score-4 anchor, though not a long enumeration of distinct actions.

4 / 5

Completeness

Explicitly states what it does (run MiniCPM5-1B in LM Studio via two runtimes with a local server) and when to use it with concrete trigger phrases ("Use when the user mentions 'LM Studio'..."), matching the score-5 anchor.

5 / 5

Trigger Term Quality

Includes natural terms a user would say — "LM Studio", desktop GUI inference, "lms" CLI, and "no-code chat UI" — giving good keyword coverage, but omits a few synonyms (e.g. "local server", "OpenAI API").

4 / 5

Distinctiveness Conflict Risk

Targets a clear niche (MiniCPM5-1B + LM Studio) with distinct triggers (LM Studio, lms CLI, desktop GUI inference), and the body's "When NOT to use" section steers away from overlapping skills, giving minimal conflict risk.

5 / 5

Total

18

/

20

Passed

Validation

93%

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

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 1 suspicious

Warning

Total

15

/

16

Passed

Repository
OpenBMB/MiniCPM
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.