Connect me to your Claude.
Query the work, not the pitch.
The endpoint is the proof: this site exposes my actual experience as agent-ready tools, so a recruiter can inspect projects, evidence, and job fit from inside their own Claude.
One URL. Six tools.
claude.ai
Open Settings → Connectors → Add custom connector, name it harsh-patel, paste the URL, and add it.
Claude Desktop
Use the same custom connector flow in Desktop. Configure the name and URL above; the remote transport is Streamable HTTP and needs no local command.
Everything reads. Nothing writes.
All six tools are READ-ONLY. The endpoint is rate-limited to 60 tool calls per hour per client.
| Tool | What it returns | Access |
|---|---|---|
about_harsh | Returns a sourced overview of Harsh's background and current work. | read-only |
search_experience | Searches Harsh's documented experience and returns linked evidence. | read-only |
get_project | Retrieves a structured brief for one shipped project. | read-only |
analyze_job_fit | Maps a job description to evidence, strengths, and honest gaps. | read-only |
get_resume | Returns Harsh's resume as structured JSON or readable Markdown. | read-only |
get_availability | Returns current location, work preferences, and availability. | read-only |
A recruiter asks. The evidence answers.
Example only — the real verdict and citations depend on the job description.
Is Harsh a fit for this platform engineering JD?
analyze_job_fit({ job_description: "…" })Strong fit. The role's Kubernetes, CI/CD, and agent-platform requirements map to shipped systems. Evidence: build-failure triage · release engineering. The report also flags requirements where the record is thin.
Put the portfolio in the workflow.
https://patelharsh.dev/api/mcp