CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

nano-banana-pro

Nano Banana/Gemini image gen/edit: text/image input, 512-4K workflows.

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

88%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 strong, operational SKILL.md: commands are verified against the actual script, the draft-to-final workflow is explicit with preflight checks and failure-recovery guidance, and there is no concept-explaining filler. Remaining improvements are minor — trimming the repeated full command invocations and splitting peripheral detail (prompt templates, test instructions) into a reference file.

DimensionReasoningScore

Conciseness

The body is dense and operational — commands, parameter mappings, error-to-fix pairs, and filename conventions — with no padding or explanations of concepts Claude already knows. Not 5 because the full "uv run ~/.codex/..." command is repeated ~5 times across Usage, Default Workflow, and Examples, and the Examples section largely duplicates Usage.

4 / 5

Actionability

Every command is copy-paste ready and its flags match the actual script's argparse (--prompt, --filename, --input-image, --resolution with 512/1K/2K/4K, --api-key); resolution request mapping, API-key precedence, preflight checks, common-failure fixes, and worked filename examples cover the common cases. Fully executable with no gaps.

5 / 5

Workflow Clarity

The draft (1K) → iterate → final (4K) workflow is explicitly sequenced with a gating rule ("Final (4K): only when prompt is locked"), and the Preflight + Common Failures section provides validation checks and an error-to-fix feedback loop for recovery. The operation is also a single unambiguous call, so the simple-skill condition holds.

5 / 5

Progressive Disclosure

Sections are clearly organized and the referenced script path (scripts/generate_image.py, plus the test file) is real and correctly documented. Not 5 because the ~130-line body is entirely inline with no reference files at all — prompt templates and filename conventions could be split out, and the test-run instruction sits oddly under "Output".

4 / 5

Total

18

/

20

Passed

Description

66%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.

The description conveys concrete capabilities (generate and edit images from text or image inputs at 512-4K) with recognizable product-name triggers, but it is entirely missing a "Use when..." trigger clause, which caps its completeness. Expanding the abbreviations and adding explicit usage triggers would make it both more complete and more naturally matchable.

Suggestions

Add an explicit trigger clause, e.g. "Use when the user asks to generate, create, or edit images, or mentions Nano Banana, Gemini image generation, or image-to-image editing."

Spell out the abbreviated actions ("generate and edit images") instead of "image gen/edit" so the description matches natural phrasing users and Claude both use for triggering.

Add one or two natural synonyms such as "picture", "photo editing", or "AI image" to broaden keyword coverage without adding bulk.

DimensionReasoningScore

Specificity

"image gen/edit" names both concrete capabilities, and "text/image input, 512-4K" specifies input modes and resolution range — several specific actions with only minor gaps. Not 5 because "512-4K workflows" is telegraphic and the actions are abbreviated rather than fully enumerated.

4 / 5

Completeness

The "what" is clear (image generation/editing with text/image inputs at 512-4K), but there is no "Use when..." clause or equivalent explicit trigger guidance, which caps completeness at 3 per the judging guidelines. Not 4 because the "when" is entirely absent rather than merely imprecise.

3 / 5

Trigger Term Quality

"Nano Banana", "Gemini", "image", "edit", "512", "4K" are natural terms users would say, giving good keyword coverage including both product nicknames. Not 5 because natural task phrases like "generate an image", "create a picture", or "photo" synonyms are missing and "gen" is abbreviated.

4 / 5

Distinctiveness Conflict Risk

Naming "Nano Banana/Gemini" establishes a clear niche with product-specific triggers, but the generic "image gen/edit" phrasing could still capture requests intended for other image tools when the user doesn't name the product — minor overlap risk. Not 5 because a broader image-generation request without the product name could route here ambiguously.

4 / 5

Total

15

/

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

steipete/agent-scripts

Path

skills/nano-banana-pro/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.