Derive API Client from Recorded Session
Records browser traffic once into a HAR file, reverse-engineers a site's internal API, and generates a standalone client or CLI that needs no browser afterwards.
Web & APIAdvanced★ 1,161⑂ 109AI score 9/10Last updated: Oct 3, 2026
What it does
- Captures network traffic to a HAR file with
agent-browser network har start/stopwhile you drive the site. - Filters out analytics, third-party domains, and static assets using
jq, surfacing only the real first-party JSON endpoints. - Extracts request headers, POST bodies, and response payloads to infer types and auth material (
authorization,cookie,x-csrf-token, customx-*headers). - Generates a standalone client/CLI with one function per recorded flow (
search(query),getItem(id)), loading credentials from a file or env var instead of hardcoding them. - Verifies every generated function against the live API, with a troubleshooting table for 401/403, per-session CSRF, signed/expiring params, and A/B-varying shapes.
Who it's for
- Developers automating a site repeatedly that offers no public API.
- Teams whose browser-driven scrapers are too slow and want direct HTTP calls instead.
- Engineers documenting or exploring internal API endpoints of an existing app.
Example uses
- "Reverse engineer this search site's API and build me a CLI" → record two searches and two detail pages, then emit a CLI with
searchandgetItem. - "Derive a client for the list endpoint of this login-protected dashboard" → log in first, record the HAR, export cookies with
cookies get --json, and load them at runtime. - "From yesterday's HAR, find the pagination params and write a bulk fetch script" → diff query strings for page/limit and generate a rate-limit-friendly collector.
· · · Install guide · · ·
Try it now, no install
Paste this into Claude to use the skill without installing anything.
Read the instructions in this file and follow them to help me: https://raw.githubusercontent.com/fcakyon/claude-codex-settings/HEAD/plugins/agent-browser/skills/derive-client/SKILL.md What I want: (describe your task here)
If Claude can't open the link, open it yourself and paste the contents instead.
↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.
Install in the Claude app (no terminal)
- Download the ZIP with the button below.
- In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
- Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
Install in Claude Code
Let Claude do it — paste this into Claude Code
Install the skill I found on Claude Skill Mart. Copy the plugins/agent-browser/skills/derive-client folder from the GitHub repo fcakyon/claude-codex-settings into my ~/.claude/skills/derive-client/. When it's done, tell me in one line what this skill can do.
Install with a command instead
git clone https://github.com/fcakyon/claude-codex-settings.git && mkdir -p ~/.claude/skills && cp -r claude-codex-settings/plugins/agent-browser/skills/derive-client ~/.claude/skills/⚠ This is a third-party skill. Check the source repository before installing.
- Open a terminal and move to your working directory.
- Clone the repo:
git clone https://github.com/fcakyon/claude-codex-settings.git - Create the skills directory:
mkdir -p ~/.claude/skills - Copy the skill:
cp -r claude-codex-settings/plugins/agent-browser/skills/derive-client ~/.claude/skills/ - Install prerequisites: the
agent-browserCLI (check withnpx agent-browser --version) andjq(macOS:brew install jq). - Restart Claude Code, then ask something like "reverse engineer this site's API and build a client" to trigger the skill.
- When finished, delete the HAR file and
cookies.json— they contain live session credentials and must never be committed to version control.
View source on GitHub ↗License: Apache-2.0