LinkedIn Profile + Job Postings Agents

# Task 003: LinkedIn Profile + Job Postings Agents

## Description
Create LinkedIn Profile Agent and Job Postings Agent following the discovery-agent.ts pattern. These are high-priority agents with highest user demand.

## Dependencies
- Depends on: [002]
- Parallel: false

## Estimated Size
L (3 days)

## Acceptance Criteria
- [ ] linkedin-profile-agent.ts created with createLinkedInProfileAgent()
- [ ] job-postings-agent.ts created with createJobPostingsAgent()
- [ ] Both export agent metadata (priority, estimatedCost, dataSource)
- [ ] LinkedIn: Extract name, title, company, profile URL, connections, skills
- [ ] Jobs: Extract active job count, hiring departments, hiring velocity
- [ ] Zod output schemas defined (LinkedInProfileResult, JobPostingsResult)
- [ ] Firecrawl fallback implemented for both agents
- [ ] Integration tests with sample inputs
- [ ] ToS disclaimer documented for LinkedIn scraping

## Implementation Notes
Follow the pattern established in discovery-agent.ts for consistency. Both agents should integrate with the existing agent architecture and support the Firecrawl fallback mechanism.

## Priority
High - These agents have the highest user demand according to requirements.

Start by reading discovery-agent.ts to understand the existing agent architecture and fallback pattern. Create linkedin-profile-agent.ts and job-postings-agent.ts with the requested metadata, Zod schemas, extraction behavior, Firecrawl fallback, and integration tests using sample inputs. Done means both agents work consistently with the pattern and the LinkedIn scraping ToS disclaimer is documented.

Written by the indexing model from the issue text.