Skip to content

v0.4.0

Choose a tag to compare

@github-actions github-actions released this 25 Feb 04:22
· 149 commits to main since this release

Changelog

  • e2f8e8a release: v0.4.0
  • bf8bcd4 docs: document arc chat, MCP tools, and async run workflow
  • 38ef8e0 feat: add stop-on-failure mode and async MCP execution for orchestrator
  • 6cec08e docs: flesh out chat-agent guide with workflow selection and plan writing guidance
  • ede8f0f feat: add MCP server and arc chat command
  • d598165 feat(dev): add re-evaluation loop to clarification step
  • ffa786e fix(monitor): widen STATE column for block-namespaced state names
  • 71bd5cf feat: add audit workflow type for adversarial-only testing
  • 6383273 feat(dev): add clarification step to arc dev pipeline
  • 9bdf5a9 feat(monitor): rich TUI with detail panel, navigation, and full phase state
  • 7ef1410 feat(dev): wire up prompt templates and add post-orchestration code review
  • 0dfc31e docs: add arc dev, composable blocks, worktrees, confidence scoring, and summaries
  • 3bf8de2 feat(arc-dev): dev command, confidence scoring, and summary generation
  • 84f0234 feat(review): confidence scoring for suggestions and plan summary generation
  • d60b09d feat(dev): arc dev command with RunDev pipeline and plan name generation
  • e084c87 feat(dev): architect agent spawning, plan generation, and custom workflow support
  • 8dc5dd6 feat(adversarial): composable block system and adversarial workflow
  • f3df069 feat(blocks): modular adversarial workflow with composable blocks
  • 9a7ec2e feat(orchestrator): worktree isolation, per-state agent config, working dir plumbing
  • 4fa1c7f feat(discovery-engine): phase: discovery-parsing
  • 367f0d0 feat(types-and-prompts): phase: core-types-and-direct
  • 0574d83 fix(orchestrator): track per-state iteration counts and sync phase status
  • f1230e9 fix(prompts): require text output, remove duplicate top-level directories
  • 0337ab0 fix(agent): unset CLAUDECODE env var when spawning subprocesses
  • 9b75c3a test(prompts): add validation suite, fix verdict branching in bugfix/refactor workflows
  • f4297d4 fix(prompts): inline partials, require text output, fix investigation verdicts
  • 73da455 chore: remove unused arc-go-pitfalls notes
  • 6a7bf77 fix(prompts): replace Rust-specific references with language-agnostic terms
  • ff34d1c chore(bench): remove integration phase plans from benchmark tasks
  • 0c4ed26 fix(prompts): remove Rust tools from analyze prompt and ensure text output
  • dc37f5c refactor(plan): remove automatic integration phase from plan creation
  • 6aaf4c8 feat(bench): merge arc-bench benchmarking suite into bench/
  • 8cfcbc7 fix(review): cap retries and enforce verdict format in review prompts
  • a984d15 docs: update build command to install to ~/.local/bin
  • fff12bc feat(agent): track token usage and cost from Claude CLI JSON output