# Arc Pro B70 Extraction

> Use this skill to extract data from Level1Techs forum threads and maintain agentic primitives in the level1techs-forums repository.

- Skill: `bojak83318/arc-pro-b70-extraction` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add bojak83318/arc-pro-b70-extraction`
- Raw SKILL.md: https://api.skillmd.com/api/skills/bojak83318/arc-pro-b70-extraction/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: bojak83318 (https://skillmd.com/u/bojak83318)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/bojak83318/arc-pro-b70-extraction

---


# Level1Techs Forum Extraction Skill

This skill automates the extraction of discussion data from various Level1Techs forum threads and maintains an organized agentic primitive archive on GitHub.

## Repository
`https://github.com/bojak83318/level1techs-forums`

## Directory Structure
Primitives are organized by category and sub-category matching the forum URL structure:
`high-performance-computing/deep-machine-learning-ai/`

## Primitive Naming
Convention: `[topic-slug]-primitive.json`

## Remote Execution
Extraction is performed within the `kasm-vm` Docker container `ff9c2ffca193` using `curl_cffi` (chrome124 impersonation) to bypass protections.

Script Location: `.gemini/skills/arc-pro-b70-extraction/scripts/extract_l1t.py`

## Workflow
1. **Prepare Script:** Ensure the topic list in `extract_l1t.py` includes the target topic IDs.
2. **Push to Container:** 
   `incus file push .gemini/skills/arc-pro-b70-extraction/scripts/extract_l1t.py kasm-vm/home/kasm-user/extract_l1t.py`
   `incus exec kasm-vm -- docker cp /home/kasm-user/extract_l1t.py ff9c2ffca193:/home/kasm-user/extract_l1t.py`
3. **Execute Extraction:**
   `incus exec kasm-vm -- docker exec -u kasm-user ff9c2ffca193 python3 /home/kasm-user/extract_l1t.py`
4. **Retrieve Primitives:**
   Pull the generated `[slug]-primitive.json` files from the container back to the host.
5. **Sync to GitHub:**
   Move primitives to the correct category folder (e.g., `high-performance-computing/deep-machine-learning-ai/`) and push to `bojak83318/level1techs-forums` using the authorized token.

## Data Schema (Agentic Primitive)
- `primitive_id`: "[topic-slug]-primitive"
- `domain`: "forum.level1techs.com"
- `source`: Full thread URL
- `data.mappings`: Array of `{username, post_number, created_at, board, topic, content_markdown}`