agent-image-skills
CLAUDE.md1 starsChanged 9 months agoMITView on GitHub
Multi-language client library monorepo for the Lab Nocturne Images API (https://images.labnocturne.com). Provides installable client packages in Python, Go, Ruby, JavaScript, and PHP, plus Claude Code skills and a ChatGPT Action integration. The API server lives elsewhere (jjenkins/labnocturne). This repo contains only client-side code and skill definitions. All five clients implement the same API surface with language-idiomatic naming: All clients default base_url to https://images.labnocturne.com and authenticate via Authorization: Bearer <key> header. Upload uses multipart/form-data. Skills are in skills/ as symlinks. Each…
- Reads credentials
- Installs packages
In this repository10 other files it gives its agents.
- deleteSkill
- filesSkill
- generate-keySkill
- image-deleteSkill
- image-filesSkill
- image-keySkill
- image-statsSkill
- image-uploadSkill
- statsSkill
- uploadSkill
The file
We split this CLAUDE.md into the rules it gives an agent
7 rules. About 0 are common, 5 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 56
All five clients implement the same API surface with language-idiomatic naming:
- Very specific: line 66
All clients default base_url to https://images.labnocturne.com and authenticate via Authorization: Bearer <key> header.
- Distinctive: line 91
Keep client code simple and beginner-friendly with minimal dependencies
- Distinctive: line 92
Use real, working code — not pseudocode
- Distinctive: line 93
Follow each language's idiomatic conventions
- Distinctive: line 94
Keep dependencies minimal (Go and Ruby clients use only stdlib)
- Very specific: line 95
Test examples against the live API before committing (/test-example <language>)
≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CLAUDE.md
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
What This Repo Is
Multi-language client library monorepo for the Lab Nocturne Images API (https://images.labnocturne.com). Provides installable client packages in Python, Go, Ruby, JavaScript, and PHP, plus Claude Code skills and a ChatGPT Action integration.
The API server lives elsewhere (jjenkins/labnocturne). This repo contains only client-side code and skill definitions.
Repository Layout
python/— Python client (labnocturnepackage, usesrequests)go/— Go client (labnocturnepackage, stdlib only)javascript/— Node.js client (ESM, usesform-data)ruby/— Ruby client (labnocturnegem, stdlibnet/http)php/— PHP client (usesext-curl, published to Packagist aslabnocturne/image-client)examples/curl/— curl-based usage examplesskills/— Claude Code skill definitions (symlinks to.agents/skills/)integrations/chatgpt-action/— ChatGPT GPT Action with OpenAPI schema.claude/commands/test-example.md— Custom command to test language examples
Build & Test Commands
Python:
pip install ./python # install locally
pip install ./python[dev] # install with dev deps (pytest, black, flake8, mypy)
cd python && pytest # run testsGo:
cd go && go build ./... # build
cd go && go test ./... # testJavaScript:
cd javascript && npm install # install deps
cd javascript && npm test # run tests (node --test)Ruby:
cd ruby && bundle install # install depsPHP:
composer install # install from root composer.jsonClient Library Conventions
All five clients implement the same API surface with language-idiomatic naming:
| Operation | Endpoint | Python | Go | JS | Ruby | PHP |
| Generate key | GET /key | generate_test_key() (static) | GenerateTestKey() | generateTestKey() (static) | generate_test_key (class) | generateTestKey() (static) |
| Upload | POST /upload | upload(path) | Upload(path) | upload(path) | upload(path) | upload(path) |
| List files | GET /files | list_files() | ListFiles() | listFiles() | list_files() | listFiles() |
| Stats | GET /stats | get_stats() | GetStats() | getStats() | get_stats | getStats() |
| Delete | DELETE /i/:id | delete_file(id) | DeleteFile(id) | deleteFile(id) | delete_file(id) | deleteFile(id) |
All clients default base_url to https://images.labnocturne.com and authenticate via Authorization: Bearer <key> header. Upload uses multipart/form-data.
Skills (Claude Code)
Skills are in skills/ as symlinks. Each skill is a SKILL.md with YAML frontmatter (name, description, metadata tags). The image-* prefixed skills are the canonical set:
image-upload— Upload an image, get CDN URLimage-files— List uploaded files with paginationimage-stats— View storage usageimage-delete— Soft-delete an imageimage-key— Generate a test API key
Skills auto-detect $LABNOCTURNE_API_KEY env var. If unset, they generate a temporary test key (7-day retention, 10MB limit). Skills use $LABNOCTURNE_BASE_URL if set.
Shared auth logic (API key resolution, base URL, common error codes) lives in .agents/skills/references/auth.md and is referenced by all skills.
The non-prefixed skills (upload, files, stats, delete, generate-key) are legacy aliases that redirect to their image-* counterparts.
API Key Types
- Test keys (
ln_test_*): free, no signup, 10MB file limit, 7-day retention - Live keys (
ln_live_*): paid, 100MB file limit, permanent storage
Style
- Keep client code simple and beginner-friendly with minimal dependencies
- Use real, working code — not pseudocode
- Follow each language's idiomatic conventions
- Keep dependencies minimal (Go and Ruby clients use only stdlib)
- Test examples against the live API before committing (
/test-example <language>)
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

