CtrlK
BlogDocsLog inGet started
Tessl Logo

using-model-endpoint

Call a registered model endpoint over its native HTTP API from the endpoint's scoped inference kernel (BASE_URL preloaded). Load once a task needs predictions from a registered model endpoint.

59

Quality

67%

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

Fix and improve this skill with Tessl

tessl review fix ./configs/microservice/bff-service/configs/agent-skills/claude-science/using-model-endpoint/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

85%

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

The body is lean, well-organized, and respectful of the context budget, clearly scoping a single direct HTTP-call task. Its main gap is actionability: it gives precise reference details but no complete executable request example.

Suggestions

Add one minimal copy-paste-ready HTTP request example (e.g. an httpx call to $BASE_URL with the Bearer header) to move actionability from guidance to fully executable.

Optionally show the compute_provider invocation with a concrete slug example so the kernel-call mechanism is unambiguous end-to-end.

DimensionReasoningScore

Conciseness

The body is dense and assumes Claude's competence: bullets like "Call the model's native API with httpx" and "don't disable it (e.g. trust_env=False)" carry information Claude would not already know, with no padded concept explanation.

3 / 3

Actionability

Gives concrete specifics (exact function signature, env names like INFER_API_KEY and HTTP_PROXY, anti-patterns) but provides no complete executable request example a reader could copy-paste, leaving the guidance incomplete relative to the level-3 anchor.

2 / 3

Workflow Clarity

The task is a single unambiguous request/response ("No job lifecycle here — direct request/response only"), and the per-rubric simple-skill note allows a score of 3 when the single action is unambiguous and no destructive/batch validation is needed.

3 / 3

Progressive Disclosure

The skill is under 50 lines with no bundle files and is organized into clear bulleted sections; per the simple-skills note that is sufficient for a top score even without external file references.

3 / 3

Total

11

/

12

Passed

Description

50%

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 specific and technically grounded but jargon-heavy and lacks an explicit, natural "Use when..." trigger. The implicit trigger guidance and overlap with a sibling skill keep it at the middle band across most dimensions.

Suggestions

Add an explicit, natural-language trigger clause such as "Use when a task needs predictions from a registered model endpoint" to lift completeness.

Soften jargon (e.g. "scoped inference kernel", "BASE_URL preloaded") in favor of terms users would naturally say, like "call a model" or "get model predictions".

Differentiate more sharply from managed-model-endpoints by naming the distinguishing trigger (native HTTP request/response vs. endpoint lifecycle management).

DimensionReasoningScore

Specificity

Names a concrete action ("Call a registered model endpoint over its native HTTP API from the endpoint's scoped inference kernel") but describes essentially one action rather than listing multiple specific concrete actions as the level-3 anchor requires.

2 / 3

Completeness

Clearly states what the skill does, but the "when" trigger ("Load once a task needs predictions from a registered model endpoint") is terse and only implied rather than an explicit "Use when..." clause, which caps completeness at 2 per the rubric guideline.

2 / 3

Trigger Term Quality

Includes some relevant terms a user might say ("model endpoint", "predictions", "HTTP API") but leans heavily on technical jargon ("scoped inference kernel", "BASE_URL preloaded") and misses common natural variations.

2 / 3

Distinctiveness Conflict Risk

The "registered model endpoint" niche is fairly specific, but it overlaps with the sibling managed-model-endpoints skill referenced in the body, so it could still trigger for the related lifecycle skill.

2 / 3

Total

8

/

12

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.

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
UnicomAI/wanwu
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.