oh-my-openagent

What it does

oh-my-openagent, formerly known in the ecosystem around oh-my-opencode, packages memory, planning, execution, verification, and orchestration patterns for coding agents.

Best fit

Codex, OpenCode, and terminal-agent users who want a stronger coding loop.

Capabilities

• Ultimate edition for OpenCode: 11 agents, lifecycle hooks, built-in MCPs, slash commands, Team Mode, ultrawork, and hashline edits.
• Codex CLI Light edition via LazyCodex: rules, comment checker, git-bash, LSP, ultrawork, ulw-loop, continuation, telemetry, and plugin-scoped MCPs.
• Team Mode can run a lead agent plus parallel members, with tmux visualization and team tools.
• Integrated code intelligence: LSP rename, goto definition, references, diagnostics, AST-Grep search and rewrite.
• Context and rules injection can load AGENTS.md and project rules into the agent at each prompt.
• Hash-anchored edits tag lines with content hashes so stale edits are rejected before corrupting files.

Use cases

• Add memory and verification loops to a coding-agent workflow.
• Coordinate planning and execution across complex repositories.
• Type ultrawork or ulw to trigger a stronger end-to-end execution loop.
• Enable Team Mode when a feature needs frontend, backend, security, and review work in parallel.
• Use /init-deep to generate layered AGENTS.md files for a large repo.
• Use LSP and AST-Grep tools for safer rename, diagnostics, references, and structural rewrites.

Source

code-yeongyu/oh-my-openagent