CtrlK
BlogDocsLog inGet started
Tessl Logo

minicpm5-deploy-litert

Run MiniCPM5-2B or MiniCPM5-1B on-device with Google's LiteRT-LM runtime — the litert-lm CLI or its OpenAI-compatible server on a desktop, the Kotlin API or the AI Edge Gallery app on Android, the same .litertlm bundle on CPU or GPU. Use when the user says "LiteRT", "LiteRT-LM", "litertlm", ".litertlm", "Android", "Edge Gallery", "on-device app", or wants one bundle for phones and desktops.

74

Quality

93%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

SKILL.md
Quality
Evals
Security

Quality

Content

86%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 tight, highly actionable deployment recipe with executable commands and code for every surface, an explicit validation step, and good progressive disclosure that pushes exhaustive Android detail to a cookbook reference. Minor room to tighten a duplicated reference link and add an explicit fix-retry loop after validation.

Suggestions

Deduplicate the docs/deployment/litert.md reference — keep it only in the Reference section — and link 'the cookbook' inline where it is first mentioned in 2B so the pointer is one level deep and unambiguous.

Turn the Validate section into a small validate→fix→retry loop (e.g. 'if the reply is not `2`, confirm BACKEND/THINKING and the FILE variant, then re-run') to lift workflow clarity to the top anchor.

Trim the one-line intro sentence that re-explains what LiteRT-LM is, since the body already shows it through commands.

DimensionReasoningScore

Conciseness

Lean and mostly free of concepts Claude already knows — commands, a Kotlin snippet, a variables table, and pitfalls carry the weight; the only slack is the docs/deployment/litert.md link stated twice (intro and Reference) and a borderline-redundant one-line intro, so it sits just below the every-token-earns-its-place anchor.

4 / 5

Actionability

Fully copy-paste ready across all surfaces: `uv tool install litert-lm`, a complete `litert-lm run` invocation with variables, an executable Kotlin Engine snippet, and a server block with a curl example, covering the common cases.

5 / 5

Workflow Clarity

Clear Install → Run (CLI/Android/server) → Validate sequence with an explicit, concrete validation step (reply contains `2`; HTTP 200; GET /v1/models lists minicpm5-2b) and a pitfalls section covering error modes, but there is no explicit validate→fix→retry loop, so it is just below the top anchor.

4 / 5

Progressive Disclosure

A clear overview with well-signaled one-level-deep references — the full Android snippet, manifest, and Galaxy S26 check are offloaded to "the cookbook" (docs/deployment/litert.md) and the server block links the server guide — keeping exhaustive detail out of the body while the core recipe stays inline.

5 / 5

Total

18

/

20

Passed

Description

100%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, specific description that covers what the skill does across multiple surfaces and exactly when to invoke it, with comprehensive natural trigger terms including the .litertlm extension. No fluff, third-person voice, no over-claims.

DimensionReasoningScore

Specificity

Names the models (MiniCPM5-2B/1B), the runtime (LiteRT-LM), and multiple concrete deployment surfaces — the litert-lm CLI, OpenAI-compatible server, Kotlin API, AI Edge Gallery app — plus the .litertlm bundle on CPU or GPU, matching the comprehensive-coverage anchor.

5 / 5

Completeness

Explicitly answers what (run MiniCPM5 on-device via LiteRT-LM across desktop/Android) and when via an explicit "Use when the user says..." clause with concrete trigger phrases, matching the top anchor.

5 / 5

Trigger Term Quality

Lists natural user phrases including synonyms and the file extension: "LiteRT", "LiteRT-LM", "litertlm", ".litertlm", "Android", "Edge Gallery", "on-device app", matching the comprehensive-synonyms-and-extensions anchor.

5 / 5

Distinctiveness Conflict Risk

Targets a specific niche (MiniCPM5 on LiteRT-LM) with highly specific triggers unlikely to fire for unrelated skills, matching the clear-niche-minimal-conflict anchor.

5 / 5

Total

20

/

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.

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 2 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.