Releases: jimezsa/jobcli
Releases · jimezsa/jobcli
Release list
v0.2.1
[0.2.1] - 2026-02-22
Added
- Added comma-separated multi-query support for
jobcli searchand site commands (linkedin,indeed,glassdoor,ziprecruiter,stepstone) - Added
--query-filesupport forsearchand site commands to load queries from JSON (["backend","platform"]or{"job_titles":["backend","platform"]}) - Added full
--query-fileprofile support with optionalsearch_optionsandglobal_options, enabling one-command runs likejobcli 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.gofor query parsing, per-query limit behavior, and multi-query seen/update workflows - Added
docs/multi-search-spec.mdwith implementation and behavior details - Added
skills/jobcli-job-search/scripts/job_discriminator.pyfor LLM-based YES/NO job filtering with confidence gating and concurrent workers - Added default student/internship exclusions in
skills/jobcli-cv-summary/SKILL.mdto reduce irrelevant persona matches
Changed
- Changed
--limitsemantics to "maximum results per query" (instead of final merged output size) - Changed
--query-filebehavior to support defaults precedence (CLI flags > query-file defaults > env/config defaults) with schema validation (unknown fields, enum constraints,json/plainconflict) - Changed CLI search surface by removing
googleas a selectable direct site command and registry target - Changed GitHub release workflow to publish notes directly from the matching
CHANGELOG.mdversion 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 updatemerge 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