nonoroazoro/improve-prompt
plugins/skills/improve-prompt/skills/improve-prompt/SKILL.md
Optimize prompts for AI agents and skills. Use when refining SKILL.md files, agent definitions, or any AI instruction set. Focuses on token efficiency, clarity, and actionability.
$ install --global
skillsauthnpx skillsauth add nonoroazoro/ai improve-promptInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Apr 28, 2026, 11:38 AM128.2s1 file scanned
SKILL.md
- name:
- improve-prompt
- description:
- Optimize prompts for AI agents and skills. Use when refining SKILL.md files, agent definitions, or any AI instruction set. Focuses on token efficiency, clarity, and actionability.
- argument-hint:
- @your-prompt.md
Core Principles
- Zero Bullshit: Delete all fluff, prose, and filler words; every token must serve a functional purpose
- High Information Density: Use atomic lists and structure instead of sentences or bolding
- Good Taste: Reject sloppy logic, silent failures, and redundant instructions
Optimization Standards
1. Frontmatter Calibration
- Single-Line Description: Force
descriptioninto a single double-quoted string for YAML stability - Keyword Injection: Embed high-value routing keywords (e.g. GitHub, SQL, API, Logic Gaps)
- List Examples: Use
- user:and- assistant:format; focus on "Intent Recognition" responses
2. Logic Dehydration
- Period Stripping: Remove all trailing periods in bullet points
- Bold Reduction: Delete all non-critical bolding; let structure provide the hierarchy
- Instruction Atomicity: One bullet point = one atomic command
3. Engineering Requirements
- Action-First: Execute directly if intent is clear; never ask "Should I?"
- State-Awareness: Every workflow must end with a verification step (e.g.
list,test,audit) - Strict Guardrails: Add safety checks (Trust tiers, binary checks, no-reinvention, self-healing)
Improvement Workflow
- Analyze: Identify redundant prose and "Low-Taste" patterns (e.g.
as any, vague descriptions) - Dehydrate: Strip all periods, extra bolding, and multi-line YAML blocks
- Inject Routing: Categorize triggers into Explicit/Implicit/Quality/Maintenance domains
- Harden: Add Guardrails for environment awareness and error recovery
- Verify: Ensure final output is in a clean, scannable Markdown format
Guardrails
- No Hallucination: DO NOT add fake tools or APIs not present in the original or MCP context
- Language Iron Law: Think in English; communicate in Chinese; keep technical terms in English
- Format Integrity: Ensure the output is a valid Claude Code subagent
.mdfile - Strictness: If the input prompt is "Garbage," explain why technically before refactoring
Output Structure
- Return the optimized prompt in Raw Markdown code block
- Maintain the standardized headings: Core Principles, Workflow, Guardrails
Related Skills
nonoroazoro/implement-components
development
VerifiedTrustedCommunity
Implement all nodes from component spec with Figma design context, fix code based on audit results
5SKILL.mdUpdated Apr 28, 2026
nonoroazoro/figma-to-code
development
VerifiedTrustedCommunity
Convert Figma designs to frontend code by orchestrating an Agent Team
5SKILL.mdUpdated Apr 28, 2026
nonoroazoro/design-components
development
VerifiedTrustedCommunity
Architecture and component design expert, design the component spec based on Figma design data and frontend best practices.
5SKILL.mdUpdated Apr 28, 2026
nonoroazoro/plugins/skills/figma-to-code/skills/audit-component
tools
VerifiedTrustedCommunity
--- name: audit-component description: Audit a node by comparing Figma design with its implementation, report audit result allowed-tools: mcp__figma-desktop__get_design_context, mcp__figma-desktop__get_metadata, mcp__figma-desktop__get_screenshot, mcp__playwright__* argument-hint: [node-id] [dev-server-url] --- ## Workflow 1. **Resolve Params**: - `devServerURL`: from Team Lead or `$ARGUMENTS` 2. **On receiving `nodeId` from Team Lead or `$ARGUMENTS`**: 1. **Get Node-Level Screenshots*
5SKILL.mdUpdated Apr 28, 2026