CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

customizing-agent-native

How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.

Quality

82%

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

77%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, rule-driven body with a clear escalation sequence, strong validation gates around the risky eject/restore operations, and mostly executable commands and code. Its weaknesses are structural: the full component contract and CLI detail are inlined in SKILL.md instead of being split into reference files, and the shared-feature composition path lacks a concrete code example.

Suggestions

Move the 17-component design-system contract enumeration and the detailed ejection CLI command reference (inspect/diff/restore) into a reference file such as references/contract.md, linked from a short summary in SKILL.md, so the body stays an overview.

Add a minimal executable example of the controller-plus-render-slot composition in "Customize A Shared Feature" — currently the only core workflow with no code or command.

Trim known-knowledge explanation such as the list of styling runtimes adapters "may use" down to the one rule that matters (className/style are optional interop hooks, not Tailwind/CVA requirements).

DimensionReasoningScore

Conciseness

The body is dense and instructional with almost no filler — e.g. "Ejection is for intentional product customization, not for hiding an upgrade failure" — but a few lines explain things Claude already knows, such as "They are ordinary React components and may use MUI-style providers, React Aria, CSS modules, CSS-in-JS, or another styling runtime." That is over-explanation of standard React styling options, keeping it below the 'every token earns its place' anchor but above the 'mostly efficient with some unnecessary explanation' anchor.

4 / 5

Actionability

Most workflows ship copy-paste-ready material: the `defineDesignSystem` registration block, the conformance-kit invocation, the four `agent-native eject` commands, and the `docs-search`/`source-search`/`rg` lookup commands. It misses the 5 anchor because "Customize A Shared Feature" — a core path — gives only abstract direction ("Prefer feature-level headless controllers over rebuilding individual widgets... Use a product-level render slot") with no controller/slot code example, and adapter implementations are named but never shown.

4 / 5

Workflow Clarity

The escalation order is an explicit six-step numbered sequence, and the risky batch operation (ejection, which copies files and rewrites imports) has genuine validation checkpoints and recovery loops: dry-run by default that prints "the file closure, consumer import rewrites, protected package contracts, and verification commands before writing", hash-gated restore that "refuses to remove locally edited files... and prints their diff instead", and an After Ejecting checklist ("Run the manifest verification commands plus the app's formatter, typecheck, and focused tests"). This matches the full validation-plus-feedback-loop-plus-checklist anchor.

5 / 5

Progressive Disclosure

The single-file body (~205 lines) has clear headers and easy navigation, but it is well over the simple-skill size where sections alone suffice, and contract-detail content that belongs in a separate reference file is inlined — notably the 17-component design-system contract enumeration ("nine leaf components... eight behavior components") and the detailed ejection CLI subcommand walkthrough. There is no reference material split out at all (no references/, scripts/, or assets/ directories exist), which is why this sits at the 'content that should be separate is inline' anchor rather than the good-structure anchor.

3 / 5

Total

16

/

20

Passed

Description

83%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 that explicitly pairs a concrete 'what' with a multi-scenario 'Use when' clause in neutral voice. The main improvement space is broader natural-synonym coverage (design system, theming, branding) and slightly sharper differentiation from the sibling docs skill.

DimensionReasoningScore

Specificity

The description names three concrete domain actions — "How to configure, compose, or eject Agent-Native features into app-owned code" — which lands on the 'several specific actions; minor gaps in coverage' anchor rather than the vague or single-action anchors. It falls short of the 5 anchor because the actions stay high-level and lack coverage of what is actually customized (design system registration, feature slots, source lookup).

4 / 5

Completeness

It explicitly answers both questions: the 'what' is "How to configure, compose, or eject Agent-Native features into app-owned code" and the 'when' is "Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source" — a concrete, multi-trigger 'Use when' clause. The voice is third-person/neutral, so no specificity penalty applies.

5 / 5

Trigger Term Quality

Natural trigger phrases users of this framework would say are well covered: "overriding shared components or integrations", "customizing a template", "adding UI to chat or headless apps", "inspecting package source". It stops short of the comprehensive anchor because a few natural synonyms — e.g. "design system", "branding", "theming", "adapting components" — are absent.

4 / 5

Distinctiveness Conflict Risk

The Agent-Native customization/ejection niche is distinct with its own trigger vocabulary ("eject", "app-owned", "headless apps"). It is not a 5 because "inspecting package source" overlaps with the related `agent-native-docs` skill's source-lookup scope, creating minor cross-trigger risk between sibling skills.

4 / 5

Total

17

/

20

Passed

Validation

81%

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

Validation — 13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

13

/

16

Passed

Repository

BuilderIO/agent-native

Path

.agents/skills/customizing-agent-native/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.