All notable changes to Flare Stack will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- 19 CLI commands:
init,spawn,destroy,status,greenlight,plan,verify,implement,audit,scan,vision,entropy,proxy,mirror,shadow,holodeck,loom,dashboard,try - Programmatic API: Full ESM exports from
src/index.tsfor agent integration - AI Executor: Multi-provider support (Google Gemini, Anthropic Claude, OpenAI, local/Ollama)
- Jira MCP Source: Primary ticket source via Model Context Protocol
- Jira REST API Source: Fallback ticket source via direct REST API
- Git Worktree Manager: Parallel branch management per ticket
- Prompt Injector: Injects JIRA context + prompt lifecycle files into worktrees
- Model Router: Intelligent model selection by task phase and tier
- Config System: Zod-validated config with GCP and Community presets
- Quality Gates: Scavenger (code analysis), Vision QA (visual regression), Entropy Hunter (mutation testing)
- Infrastructure: Proxy Router (per-worktree dev servers), Context Holodeck (save/restore), Production Mirror (BigQuery logs)
- Extras: Shadow Load Tester (stress testing), Loom Generator (demo recording)
- Dashboard: Ink TUI + chalk fallback for real-time status
- Agent Workflow:
.agent/workflows/flare.mdfor Antigravity IDE integration - CI/CD: GitHub Actions pipeline with Node 20/22 matrix, TypeScript check, Prettier lint
- Test Suite: 18 test files, 114 tests covering all modules
- Project Hygiene:
.editorconfig,.prettierrc,CHANGELOG.md
- ESM-only TypeScript project
- Dual-mode: CLI tool + programmatic library
- 4-phase AI pipeline: Plan → Verify → Implement → Audit
- MCP-first Jira integration with REST API fallback