nano-banana-pro
Generate/edit images with Nano Banana Pro (Gemini 3 Pro Image). Use for image create/modify requests incl. edits. Supports text-to-image + image-to-image; 1K/2K/4K; use --input-image.
92
88%
Does it follow best practices?
Impact
100%
3.12xAverage score across 3 eval scenarios
Risky
Do not use without reviewing
Quality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a strong, concise description that packs specific capabilities, trigger terms, and usage guidance into a compact format. It names the specific tool/model, lists concrete actions (generate, edit, text-to-image, image-to-image), includes resolution options, and provides a clear 'Use for...' clause. The abbreviations (incl., 1K/2K/4K) keep it terse without sacrificing clarity.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple concrete actions: generate images, edit images, text-to-image, image-to-image, supports multiple resolutions (1K/2K/4K), and mentions the --input-image flag. These are specific, actionable capabilities. | 3 / 3 |
Completeness | Clearly answers 'what' (generate/edit images with specific model, text-to-image, image-to-image, multiple resolutions) and 'when' ('Use for image create/modify requests incl. edits'). The 'Use for...' clause serves as an explicit trigger guidance. | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms: 'image', 'generate', 'edit', 'create', 'modify', 'text-to-image', 'image-to-image', and the specific model names 'Nano Banana Pro' and 'Gemini 3 Pro Image'. Users asking about image generation or editing would naturally use these terms. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive due to naming the specific model (Nano Banana Pro / Gemini 3 Pro Image) and specifying the exact capabilities and flag (--input-image). This would clearly differentiate from other image or general AI skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a solid, actionable skill with clear executable commands and a well-defined iterative workflow. Its main weakness is moderate verbosity—several sections (filename examples, resolution mapping, prompt templates) could be tightened without losing clarity. The progressive disclosure could benefit from splitting detailed reference material into separate files.
Suggestions
Condense the filename generation section to just the format pattern and 1-2 examples instead of 4
Consider moving prompt templates and resolution mapping details to a separate reference file, keeping only the essentials inline
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is mostly efficient but includes some sections that could be tightened. The filename generation section with multiple examples is somewhat verbose, the resolution mapping table over-explains, and the prompt templates section adds bulk. The preflight section and API key order are useful but could be more compact. | 2 / 3 |
Actionability | Fully executable commands are provided throughout with concrete examples. The bash commands are copy-paste ready with clear parameter flags, and the workflow steps give specific resolution flags and filename patterns. Both generation and editing use cases have complete, runnable examples. | 3 / 3 |
Workflow Clarity | The draft → iterate → final workflow is clearly sequenced with explicit resolution progression (1K → 4K). The preflight checks serve as validation steps, common failures provide error recovery guidance, and the editing workflow has clear steps. The iterative feedback loop (draft, adjust prompt, finalize) is well-defined. | 3 / 3 |
Progressive Disclosure | The content is well-organized with clear section headers, but it's a fairly long monolithic document. Some sections like prompt templates, filename generation details, and resolution mapping could potentially be split into referenced files. However, there are no deeply nested references, and the structure is navigable. | 2 / 3 |
Total | 10 / 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.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- Dicklesworthstone/pi_agent_rust
- Commit
6e3d68c
- 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.