v0.2.1

[0.2.1] - 2026-02-22

Added

• Added comma-separated multi-query support for jobcli search and site commands (linkedin, indeed, glassdoor, ziprecruiter, stepstone)
• Added --query-file support for search and site commands to load queries from JSON (["backend","platform"] or {"job_titles":["backend","platform"]})
• Added full --query-file profile support with optional search_options and global_options, enabling one-command runs like jobcli search --query-file queries.json
• Added multi-query parsing/validation: split and trim tokens, remove empties, case-insensitive dedupe, empty-query error, and max 10 queries
• Added cross-query merge/dedupe flow using seen-key normalization (title + company, fallback URL) plus per-query limiting before merge
• Added regression coverage in internal/cmd/search_test.go for query parsing, per-query limit behavior, and multi-query seen/update workflows
• Added docs/multi-search-spec.md with implementation and behavior details
• Added skills/jobcli-job-search/scripts/job_discriminator.py for LLM-based YES/NO job filtering with confidence gating and concurrent workers
• Added default student/internship exclusions in skills/jobcli-cv-summary/SKILL.md to reduce irrelevant persona matches

Changed

• Changed --limit semantics to "maximum results per query" (instead of final merged output size)
• Changed --query-file behavior to support defaults precedence (CLI flags > query-file defaults > env/config defaults) with schema validation (unknown fields, enum constraints, json/plain conflict)
• Changed CLI search surface by removing google as a selectable direct site command and registry target
• Changed GitHub release workflow to publish notes directly from the matching CHANGELOG.md version section
• Updated command overview examples and docs (README.md, docs/usage.md) including seen-update flow, multi-query examples, and expanded flag tables
• Updated skills docs and search flow to use a binary persona filter pipeline with jobcli seen update merge steps and enumerated output formatting
• Updated job discriminator defaults/behavior (model/API setup, timeout/tokens/workers, progress logging, and more permissive matching guidance)

Fixed

• Fixed environment loading in the pirate motivator script by reading project-root .env
• Fixed job discriminator throughput by increasing default parallel worker count for faster batch processing