Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
95 lines (73 loc) · 3.39 KB

CONTRIBUTING.md

File metadata and controls

95 lines (73 loc) · 3.39 KB

Contributing to César

Thanks for your interest in fighting parasitic lead magnets on LinkedIn! Here's how to contribute.

Quick start

git clone https://github.com/bastiensoret/Cesar.git
cd Cesar
npm install
npm run build
  1. Open Chrome → chrome://extensions/
  2. Enable Developer mode
  3. Click Load unpacked → select the dist/ folder
  4. Go to linkedin.com and scroll your feed

How to contribute

Report a bug or false positive

Open an issue with:

  • Screenshot of the badge (or lack thereof)
  • Screenshot of the console logs (F12, filter on [César])
  • Your verdict: was César right or wrong?
  • The post URL if possible

Add regex patterns

The detection patterns live in src/content/ in the THIRD_PARTY_PATTERNS and GATING_CTA_PATTERNS arrays. To add a new pattern:

  1. Identify a recurring CTA or third-party attribution phrase
  2. Write a regex that matches it (test on regex101.com first)
  3. Add it to the appropriate array with an English label and appropriate weight
  4. Test with __cesar.test("your test text") in the console
  5. Submit a PR with before/after screenshots

Fix LinkedIn DOM selectors

LinkedIn changes their DOM structure frequently. If César stops detecting posts:

  1. Open DevTools on LinkedIn
  2. Inspect the post structure and find the new selectors
  3. Update the selectors in src/content/ (search for querySelector)
  4. Test that scanning works again
  5. Submit a PR

Improve the LLM prompt

The system prompt lives in src/background/ in the SYSTEM_PROMPT constant. If you find a false positive or false negative with AI enabled:

  1. Share the post text and the LLM's response
  2. Suggest a prompt modification that would fix it
  3. Test with at least 5 other posts to make sure it doesn't break existing cases

Add a new language

Currently César detects FR and EN patterns. To add a new language:

  1. Add patterns to THIRD_PARTY_PATTERNS (how people reference third-party content in that language)
  2. Add patterns to GATING_CTA_PATTERNS (how people gate content in that language)
  3. Test on real LinkedIn posts in that language
  4. Update the README with the new supported language

Code structure

Directory Purpose
src/content/ Content script — detection engine + overlay injection
src/background/ Service worker — LLM API routing + prompt caching
src/popup/ Extension popup UI (HTML + CSS + JS)
src/debug/ MAIN world bridge for console debugging
src/shared/ Shared utilities (storage, sanitization)
static/ Icons and CSS copied to dist/ as-is
tests/ Vitest unit tests with jsdom

Coding conventions

  • All static UI text in English
  • LLM-generated content adapts to post language
  • Pattern labels in English
  • Console logs prefixed with [César]
  • No external runtime dependencies — fetch-only for LLM calls
  • Run npm run lint and npm run test before committing

Pull request process

  1. Fork the repo and create a feature branch (git checkout -b feature/my-feature)
  2. Make your changes
  3. Test on your LinkedIn feed with debug mode on
  4. Commit with a clear message describing what changed
  5. Open a PR with:
    • What you changed and why
    • Screenshots if UI-related
    • Test results (posts tested, verdicts)

Code of conduct

See CODE_OF_CONDUCT.md. Be respectful, constructive, and focused on making LinkedIn a better place.