Skip to content

v2.3.0

Choose a tag to compare

View all tags
@github-actions github-actions released this 18 Dec 19:13
· 46 commits to dev since this release
v2.3.0
7b54c2a

v2.3.0 - Meet Sisyphus

Introducing Our Main Agent: Sisyphus

Meet Sisyphus

In Greek mythology, Sisyphus was condemned to roll a boulder up a hill for eternity as punishment for deceiving the gods. LLM Agents haven't really done anything wrong, yet they too roll their "stones"—their thoughts—every single day. My life is no different. Looking back, we are not so different from these agents.

Yes! LLM Agents are no different from us. They can write code as brilliant as ours and work just as excellently—if you give them great tools and solid teammates.

I believe that code written by agents should be indistinguishable from human-written code. This philosophy inspired the name "Sisyphus"—an agent that rolls its boulder every day, just like us, producing work that seamlessly blends with human output.

Sisyphus's Toolkit

Meet our main agent: Sisyphus (Claude Opus 4.5 with 32k extended thinking). Below are the tools Sisyphus uses to keep that boulder rolling:

  • Curated Teammates
    • Oracle: Design & debugging (GPT 5.2)
    • Frontend UI/UX Engineer: Beautiful interfaces (Gemini 3 Pro)
    • Librarian: Official docs & OSS implementations (Claude Sonnet 4.5)
    • Explore: Blazing fast codebase exploration (Grok Code)
  • Full LSP / AST-Grep Support: Refactor decisively
  • Todo Continuation Enforcer: Forces the agent to continue if it quits halfway—this is what keeps Sisyphus rolling that boulder
  • Comment Checker: Prevents AI from adding excessive comments. Code generated by Sisyphus should be indistinguishable from human-written code

Enhanced Orchestration

Sisyphus brings more refined orchestration capabilities:

  • Smarter delegation to specialized subagents
  • Aggressive parallel execution for maximum throughput
  • Todo-driven workflow with real-time progress tracking
  • Prevented proactive implementation without explicit user request (#106)

Experimental Features

New opt-in experimental features for power users. Use with caution—these may change or be removed in future versions.

{
  "experimental": {
    "aggressive_truncation": true,
    "empty_message_recovery": true,
    "auto_resume": true
  }
}
Option Description
aggressive_truncation When token limit is exceeded, aggressively truncates tool outputs to fit within limits
empty_message_recovery Automatically recovers from "non-empty content" API errors by fixing empty messages
auto_resume Automatically resumes session after successful recovery from thinking block errors

Other Changes

  • Added Simplified Chinese README (zh-cn)
  • Fixed circular dependency in config error utilities preventing plugin loader crash
  • Added Windows config path documentation
  • Fixed background-agent to cancel all nested descendant tasks recursively (#107)
  • Increased todo-continuation-enforcer delay to 5s with write permission check (#105)

Commits

  • df87f5f Introducing our main agent: Sisyphus (#113)
  • 4cd2745 refactor(auto-update-checker): remove config path from startup toast
  • 8cf713e feat(config): add experimental config for gating unstable features (#110)
  • 7fe6423 docs: add Simplified Chinese README (zh-cn)
  • dad534e fix: break circular dependency in config error utilities to prevent plugin loader crash
  • 63fea77 fix: add Windows config path documentation and config error warnings (#97) (#109)
  • 845a1d2 fix(background-agent): cancel all nested descendant tasks recursively (#107)
  • df0a9e6 Prevent OmO from proactively implementing without explicit user request (#106)
  • a48fc3e fix(todo-continuation-enforcer): increase delay to 5s and add write permission check (#89) (#105)
Assets 2
Loading