How to Enable Specific Job Portal Skills in the AI-Job-Search Scraper

Set enabled: true in the frontmatter of any portal's SKILL.md file located under .agents/skills/[portal-name]/ to include that job source in the next /scrape execution.

The MadsLorentzen/ai-job-search repository uses a modular skill system where each job portal is defined by a dedicated SKILL.md file. The scraper automatically discovers these portals by scanning the .agents/skills/*/SKILL.md directory structure and respects an enabled toggle in each file's YAML frontmatter to determine which sources to query during a run.

Understanding the Skill Discovery Mechanism

According to the job-scraper skill specification in .claude/skills/job-scraper/SKILL.md, the scraper performs automatic discovery at runtime. It reads every SKILL.md found under .agents/skills/*/SKILL.md and inspects the frontmatter for an enabled flag.

Portals are enabled by default if the key is omitted. Only when enabled: false is explicitly set does the scraper skip that portal and report it as "skipped (disabled)" in the output. This design allows you to maintain a library of portal integrations while activating only the relevant ones for specific searches.

Step-by-Step: Enabling a Job Portal

Locate the Portal's SKILL.md

Each job portal lives in its own subdirectory under .agents/skills/. For example:

• LinkedIn: .agents/skills/linkedin-search/SKILL.md
• Jobbank: .agents/skills/jobbank-search/SKILL.md
• Freehire: .agents/skills/freehire-search/SKILL.md

Edit the Frontmatter Enabled Flag

Open the target SKILL.md and modify the YAML frontmatter at the top of the file. Change the enabled value to true (or add the key if missing):

---
name: jobbank-search
version: 1.0.0
description: Search live Danish job listings from Jobbank
enabled: true   # Set to true to include, false to skip

---

If you need to disable a portal temporarily without deleting its configuration, set enabled: false:

sed -i 's/enabled: true/enabled: false/' .agents/skills/freehire-search/SKILL.md

Execute the Scraper

Run the scraper command to apply your changes:

claude code /scrape

The scraper will now include the newly enabled portal in its parallel search batch, using the CLI command defined in that portal's SKILL.md. Disabled portals will be explicitly listed as skipped in the execution report.

Bulk Configuration Strategies

To enable multiple portals simultaneously, you can script the modification across several files. Since each portal operates independently, there is no central configuration file—each SKILL.md controls its own state. This modular approach ensures that changes to one portal's availability do not affect others.

For example, to enable all portals in a directory:

for skill in .agents/skills/*/SKILL.md; do
  sed -i 's/enabled: false/enabled: true/' "$skill"
done

Summary

• The scraper discovers portals by reading .agents/skills/*/SKILL.md files automatically.
• The enabled flag in YAML frontmatter controls inclusion; omitting the key defaults to enabled.
• Set enabled: true to activate a portal, or enabled: false to skip it without removing code.
• Changes take effect immediately on the next /scrape invocation—no service restart or additional wiring required.
• Discovery and filtering logic is implemented in .claude/skills/job-scraper/SKILL.md.

Frequently Asked Questions

What happens if the enabled key is missing from SKILL.md?

If the enabled key is absent from the frontmatter, the portal is treated as enabled by default. The scraper will include it in the search batch unless explicitly disabled with enabled: false.

Can I enable multiple portals at once?

Yes. Since each portal has its own SKILL.md, you can edit multiple files to set enabled: true before running the scraper. The /scrape command will execute all enabled portals in parallel, respecting each file's individual toggle.

Where is the scraper discovery logic defined?

The discovery mechanism and enabled flag handling are documented in .claude/skills/job-scraper/SKILL.md at lines 60-72. This specification defines how the scraper reads the frontmatter and decides which portals to execute.

Do I need to restart the CLI after changing enabled flags?

No. The scraper reads the SKILL.md files fresh during each /scrape invocation. Changes to the enabled field are picked up immediately on the next run without requiring a restart or reload of the environment.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}