-
Notifications
You must be signed in to change notification settings - Fork 4
Home
JobSentinel is an open-source, local-first desktop job-search assistant for finding real, relevant, fairly compensated work while keeping sensitive job-search data under user control.
This wiki is public product documentation. The repository docs remain the source of truth for implementation detail, verification, and release work.
| Topic | Current state |
|---|---|
| Source version |
2.9.0 in main; local pre-release gates pass |
| Latest published macOS package | Legacy v2.7.7 no-account universal DMG with matching .sha256
|
| Latest full cross-platform release | Legacy v2.7.5 Windows, macOS, and Linux assets |
| Pending public assets | Fresh public Windows MSI/setup EXE, Linux, macOS, SBOM, attestation, and Agent Skills assets are still pending for 2.9.0
|
| macOS readiness | Full-public readiness is 94%; no-account path completion is 100% |
| App model | Desktop app built with React 19, TypeScript, Tauri 2, Rust, and SQLite |
| Data model | Local-first by default, with no hosted account or telemetry requirement |
| External AI | Optional, disabled by default, preview-gated, and routed through one privacy-first AI gateway |
- Current Capabilities - maintained summary of active product capabilities, boundaries, and release posture.
- README - public project front door.
- Quick Start - install and first-run guide.
- Privacy - local-first data boundaries.
- Responsible AI - AI use and misuse boundaries.
- Architecture - technical system map.
- Roadmap - current priorities and limits.
- Active Plan Status - current restart surface for active work.
User privacy and security are non-negotiable. JobSentinel must preserve local-first storage, credential safety, source boundaries, explicit user review, and privacy-preserving defaults.
Core workflows work locally. Job tracking, saved searches, application notes, salary floors, resume records, source status, and safe support reports work without a hosted account, telemetry, cloud sync, or external AI provider.
The 2.9.0 no-account release path can publish both Windows MSI and setup EXE
assets. If Windows signing credentials are unavailable, those installer
filenames must include _unsigned and each installer must have a matching
.sha256 checksum. Windows SmartScreen warnings are expected for unsigned or
new, low-reputation installers.
The current local 2.9.0 no-account universal DMG has been rebuilt from current
source and verified with checksum, metadata, universal architecture, signature,
mounted-app launch, installed-app launch, visible-window smoke, and private
isolated-data checks. Public 2.9.0 macOS assets are not published yet.
Until the project has Apple Developer Program materials, public macOS packages
must stay clearly labeled as no-account packages, include a matching
.dmg.sha256 checksum, and avoid any Gatekeeper-ready claim. First open may
require approval in macOS Privacy & Security. Zero-friction public distribution
still needs Developer ID signing, notarization, stapling, and signed
public-artifact verification.
The 2.9.0 release path includes downloadable Agent Skills archives for job
search planning, posting-risk review, resume tailoring, form review,
application tracking, outreach, interview prep, and offer/pay review.
The release should attach both archive formats:
JobSentinel-2.9.0-agent-skills.tar.gzJobSentinel-2.9.0-agent-skills.zip
Each archive needs a matching .sha256 sidecar. Use ZIP on Windows when the
built-in extractor is preferred.
- Start with the Quick Start.
- Search existing issues.
- Open a new issue.
- Use the in-app safe support report before sharing problem details.
Last updated: June 19, 2026.