CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

self-customize

Customize your own agent — add capabilities, install packages, add MCP servers, edit code or CLAUDE.md. Use when the user asks you to add a feature, install a tool, or modify how you work. For non-trivial code changes, delegate to a builder agent via create_agent.

Quality

84%

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

82%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-executed instruction-only skill: concrete tool-call templates, clear decision routing, and worked examples that cover the common customization paths. The only real gaps are the trimmable rationale section, conditional rather than explicit validation in the main workflow, and the inline reusable builder-instruction block.

DimensionReasoningScore

Conciseness

The body is dense and assumes competence — a compact decision tree, copy-paste tool calls, and worked examples with little padding. It is not a 5 because the 'Diff Size Limits — Why' section spends ~90 words justifying a policy, and phrases like 'This gives the user a reviewable boundary' could be trimmed; it is well above anchor 3 since almost everything is actionable rather than explanatory.

4 / 5

Actionability

Fully executable, copy-paste-ready commands throughout: `create_agent({ name: "Builder", instructions: ... })`, `send_to_agent({ agentGroupId, text: ... })`, `install_packages({ apt: ["ffmpeg"], npm: ["@xenova/transformers"], reason: ... })`, and `add_mcp_server({ name: "rss", command: "npx", args: [...] })`, with two worked examples covering the common cases (adding an MCP tool, installing a system tool).

5 / 5

Workflow Clarity

The builder workflow is a clear numbered sequence (describe → create → send → review → confirm) and the builder instructions include checkpoints ('Test if possible', 'If a change would break existing tests, stop and report', 'No silent failures'). It falls short of 5 because validation is delegated and conditional rather than an explicit validate-and-retry loop in the main workflow.

4 / 5

Progressive Disclosure

Well-organized single-file structure with clear section headers (decision tree, workflow, examples, when-not-to) and no nested or dead references. It is a 4 rather than 5 because the ~30-line builder-instructions block is reusable content that could live in a reference file the skill points to, and the body exceeds the under-50-lines simple-skill exception.

4 / 5

Total

17

/

20

Passed

Description

78%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 with explicit what/when structure, concrete capability enumeration, and natural trigger phrases. Its main weakness is the second-person voice ('asks you to add a feature'), which the rubric penalizes on specificity, plus modest synonym coverage for trigger terms.

Suggestions

Rewrite in third person to avoid the specificity penalty: 'Customizes the agent by adding capabilities, installing packages, adding MCP servers, and editing code or CLAUDE.md. Use when the user asks to add a feature, install a tool, add an MCP server, or modify how the agent works.'

Broaden trigger-term coverage with synonyms users would naturally say: 'customize', 'set up', 'configure', 'add a tool/server', alongside the existing 'add a feature, install a tool'.

DimensionReasoningScore

Specificity

Names several concrete actions ('install packages, add MCP servers, edit code or CLAUDE.md', 'delegate to a builder agent via create_agent'), which sits at the anchor-4 level, but the explicit second-person phrasing ('Use when the user asks you to add a feature') triggers the rubric's one-point voice penalty, dropping it to 3. It is not a 2 because the domain and actions are concrete, not generic.

3 / 5

Completeness

Explicitly answers both what ('Customize your own agent — add capabilities, install packages, add MCP servers, edit code or CLAUDE.md') and when ('Use when the user asks you to add a feature, install a tool, or modify how you work') with concrete trigger phrases, matching the anchor-5 example structure.

5 / 5

Trigger Term Quality

'Use when the user asks you to add a feature, install a tool, or modify how you work' contains natural phrases a user would actually say. It is a 4 rather than 5 because synonym coverage is thin — 'customize', 'set up', or explicit 'MCP'/'package' trigger words are absent or underrepresented.

4 / 5

Distinctiveness Conflict Risk

The self-modification niche ('Customize your own agent', 'add MCP servers', 'delegate to a builder agent via create_agent') is clearly distinct from general coding or document skills. Minor overlap risk remains with generic code-editing skills via 'edit code', keeping it below 5.

4 / 5

Total

16

/

20

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.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

nanocoai/nanoclaw

Path

container/skills/self-customize/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.