customize

Create, configure, and maintain custom agent profiles and author new skills via the `repl` tool. Use when the user wants to create an agent profile, build a custom agent, modify agent capabilities, attach or detach skills/connectors on a profile, author a skill, or inspect which connectors and tools are available. Also use whenever you need the `host.agents.*` or `host.skills.*` Python SDK.

Quality

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

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 highly actionable, well-sequenced skill body with excellent executable code and validation checkpoints. Its main weakness is repetition and density: several concepts are explained multiple times and large reference material lives inline rather than in bundled files.

Suggestions

De-duplicate the unrestricted-vs-curated connector and excludedTools pattern-matching explanations — state them once in the SDK reference and reference them from the workflow.

Move the long host.agents/host.skills API reference and the kernel-sidecar validation rules into separate bundled reference files, keeping SKILL.md an overview with one-level-deep pointers.

Trim inferable explanations (e.g. restating why `__file__` is unset) to lean on Claude's competence and reduce token cost.

Dimension	Reasoning	Score
Conciseness	The body is dense and mostly earns its place with host-specific gotchas, but it repeats the unrestricted-vs-curated / excludedTools / pattern-matching explanations across the SDK reference, the workflow, and step 6, and includes some over-explanation a competent Claude could infer.	3 / 5
Actionability	Provides fully executable, copy-paste-ready code blocks for every SDK call, including exact signatures, return shapes, and concrete patterns like the safe skill_names update check and the co_filename sidecar example.	5 / 5
Workflow Clarity	The scope→draft→review→create workflow is explicitly sequenced with an `ask_user` review gate, read-back verification steps, and destructive-operation guidance (delete approval card behavior, exact-replace warnings, sidecar gate probe iteration).	5 / 5
Progressive Disclosure	Well-structured with clear sections and a deliberate hand-off to skill-creator for the full authoring guide, but the kernel sidecar material and the full host.agents/host.skills reference are inlined in SKILL.md rather than split into separate one-level-deep reference files.	4 / 5
	Total	17 / 20 Passed

Description

88%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, comprehensive description with explicit what-and-when guidance and many concrete actions. Trigger terms are natural but slightly tool-name-anchored, leaving minor synonym gaps.

Suggestions

Add a few natural-language synonyms users might say (e.g. 'add a skill', 'set up an agent', 'wire up a connector') alongside the SDK-name triggers.

Consider tightening 'inspect which connectors and tools are available' so it reads less like a generic capability and more like a distinct trigger.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'create, configure, and maintain custom agent profiles', 'author new skills', 'modify agent capabilities', 'attach or detach skills/connectors', 'inspect which connectors and tools are available' — covering the skill comprehensively.	5 / 5
Completeness	Explicitly states both what it does ('Create, configure, and maintain custom agent profiles and author new skills') and when to use it via two 'Use when' / 'Also use whenever' clauses with concrete trigger scenarios.	5 / 5
Trigger Term Quality	Includes natural user-facing phrases like 'create an agent profile', 'build a custom agent', 'author a skill', plus the SDK triggers 'host.agents.' / 'host.skills', but lacks synonyms and is anchored to the `repl` tool name rather than broader phrasings a user might say.	4 / 5
Distinctiveness Conflict Risk	The agent-profile/skill-authoring niche tied to `host.agents.`/`host.skills.` is fairly distinct, but 'modify agent capabilities' and 'inspect connectors and tools' have some overlap with general agent-management skills.	4 / 5
	Total	18 / 20 Passed

Validation

93%

Warnings & errors only

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

Criteria	Description	Result
referenced_paths_exist	Referenced path issues: 1 missing	Warning

	Total	15 / 16 Passed

Repository: UnicomAI/wanwu
Path: configs/microservice/bff-service/configs/agent-skills/claude-science/customize/SKILL.md
Commit: 99df6e4

Reviewed: 3 days ago

Table of Contents

Discovery Implementation Validation

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.