AI-powered code review agent that analyzes GitHub repositories to verify whether a candidate’s resume accurately reflects their actual projects, skills, and development behavior.
- Copy or create a project-root
.envwith your API keys (see app startup checks for required variables). - Optional: install GitNexus globally for richer graphs:
npm install -g gitnexus. If GitNexus is missing or fails, CodeLens falls back to a direct git clone and filesystem scan.
- To enable login, create a GitHub OAuth App at
github.com/settings/developerswith callback URL set to{APP_URL}/component/streamlit_oauth.callbackor your Streamlit redirect URI.
[x] CodeLens logo [x] Match cluely.com design for inspo [x] Add list of workflow steps [x] Popups with speedometers [] Make resume analysis simpler [x] Add interactive timeline of commits [x] Add code graph (interactive graph that looks like gitnexus codebase graph or obsidian knowledge to look like) [x] Deploy on Repl.it and/or Vercel
- Overall Quality, AI Usage, Commit Health, etc. to Github style cards (see Latest from our changelog on Github.com)
- Bullet point descriptions for each of the 5 main scores
- Visually show the scores with a spedometer look
- Clicking speedometer will open up a scrollable pop-up to corresponding description + graph
- Proper dividers between description sections
Commit Health:
- Consider timeline of commits (for commit health pop up)
Resume:
- Save resume for easy access later (if possible)
- Click Resume match box to pop up resume with color coded highlights over what matches vs what does not match
- Annotated descriptions pointing to color coded highlighted parts (red, yellow, green)
Job Fit:
- Description will also have github style bullet points (see Latest from our changelog on Github.com)
Overall order:
- 5 result speedometer cards
- Strengths/concerns
- Graph
- Skill map
- Recommendations