Skip to content

VERIFY.md

Latest commit

 

History

History
executable file
·
152 lines (104 loc) · 4.45 KB

VERIFY.md

File metadata and controls

executable file
·
152 lines (104 loc) · 4.45 KB

BrightData Skill v2.3.0 - Verification

This verification guide is for AI agents to confirm successful installation.

MANDATORY COMPLETION CHECKLIST

AI Agents: You MUST verify ALL items below pass before reporting installation success.

Core Installation (Checks 1-5)

# Check Command Pass Criteria Status
1 Directory exists ls -la $PAI_DIR/skills/BrightData/ Directory with files shown
2 SKILL.md present cat $PAI_DIR/skills/BrightData/SKILL.md | head -5 Shows frontmatter with name: BrightData
3 Workflow directory ls $PAI_DIR/skills/BrightData/Workflows/ Shows FourTierScrape.md
4 Workflow content cat $PAI_DIR/skills/BrightData/Workflows/FourTierScrape.md | head -10 Shows workflow header
5 Version correct grep "version:" $PAI_DIR/skills/BrightData/SKILL.md Shows 2.3.0

Tier 1 Verification (Checks 6-7)

# Check Command Pass Criteria Status
6 WebFetch available N/A (built-in tool) WebFetch tool exists in Claude Code
7 WebFetch works Use WebFetch on https://example.com Returns page content

Tier 2 Verification (Checks 8-9)

# Check Command Pass Criteria Status
8 Curl available which curl Returns path to curl
9 Curl with headers See curl command below Returns HTML content

Tier 2 test command:

curl -s -L -A "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36" \
  -H "Accept: text/html" \
  "https://example.com" | head -20

Tier 3 Verification (Checks 10-11) - OPTIONAL

# Check Command Pass Criteria Status
10 Browser skill installed ls $PAI_DIR/skills/Browser/SKILL.md File exists
11 Browser works bun run $PAI_DIR/skills/Browser/Tools/Browse.ts https://example.com Returns screenshot + diagnostics

Note: Tier 3 checks are OPTIONAL. Skip if Browser skill is not installed.

Tier 4 Verification (Checks 12-13) - OPTIONAL

# Check Command Pass Criteria Status
12 Bright Data MCP configured Check claude_desktop_config.json brightdata server present
13 Bright Data works mcp__Brightdata__scrape_as_markdown Returns content

Note: Tier 4 checks are OPTIONAL. Skip if Bright Data MCP is not configured.

Completion Requirements

Minimum for success:

  • Checks 1-9 must ALL pass
  • Checks 10-13 are optional based on user's tier configuration

Report format:

INSTALLATION VERIFIED: [X/Y] checks passed
   - BrightData skill v2.3.0 installed at $PAI_DIR/skills/BrightData/
   - Tier 1 (WebFetch): Ready
   - Tier 2 (Curl): Ready
   - Tier 3 (Browser): [Ready / Not configured]
   - Tier 4 (Bright Data): [Ready / Not configured]

Quick Verification

1. Check Installation

ls -la $PAI_DIR/skills/BrightData/

Expected: Directory with SKILL.md, README.md, Workflows/

2. Test Tier 1 (WebFetch)

Ask Claude:

Use WebFetch to get content from https://example.com

Expected: Returns page content as markdown

3. Test Tier 2 (Curl)

curl -s -L -A "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36" \
  -H "Accept: text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/apng,*/*;q=0.8" \
  -H "Accept-Language: en-US,en;q=0.9" \
  "https://example.com" | head -30

Expected: Returns HTML content

4. Test Full Workflow

Ask Claude:

Scrape https://example.com using the BrightData skill

Expected: Returns content with tier information

Common Issues

"Skill not found"

Verify skill directory:

ls $PAI_DIR/skills/BrightData/

"Tier 3 failing"

Install Browser skill:

# Install pai-browser-skill pack

"Tier 4 failing"

Configure Bright Data MCP in claude_desktop_config.json

Full Test Sequence

  1. Tier 1: Use WebFetch on https://example.com
  2. Tier 2: Run curl command above
  3. Tier 3 (if available): Use Browser skill on a JavaScript site
  4. Tier 4 (if available): Use Bright Data MCP on a protected site
  5. Full workflow: Ask to scrape a URL and verify automatic tier selection

All tiers should return content in markdown format.