From oh-my-agent
Runs strict Ralph mode quality gate orchestration enforcing plan-prd-exec-verify-fix pipeline with mandatory verification loops. Outputs gate status table, unresolved risks, and release decision.
npx claudepluginhub sswym/oh-my-agent --plugin oh-my-agent运行 oma `ralph` 模式。 目标： $ARGUMENTS Ralph 协议： 1. 强制执行完整的分阶段管道：`team-plan -> team-prd -> team-exec -> team-verify -> team-fix`。 2. 从不跳过 `team-verify`。 3. 如果验证失败，运行 `team-fix` 然后返回到 `team-verify`。 4. 不要在没有所有验收标准的通过证据的情况下声明完成。 5. 明确升级未解决的阻塞。 模式姿态： - 优先规划/审查深度。 - 倾向于保守、可逆的决策。 - 保持最终输出以发布为导向。 输出格式： ## 模式 - ralph ## 门控状态 | 门控 | 状态 | 证据 | | --- | --- | --- | | team-plan | ... | ... | | team-prd | ... | ... | | team-exec | ... | ... | | team-verify | ... | ... | | team-fix | ... | ... | ## 未解决风险 - ... ## 发布决策 - 就绪 / 未就绪
/ralphInitiates persistent self-referential loop by reading and following instructions in skills/ralph/SKILL.md until verification passes.
/ralphRuns Ralph AI automation toolkit portably on any git repository with workspace targeting. Supports run (multi-iteration via /loop), status, dashboard subcommands and LLM tools.
/ralphRuns autonomous development loops with fresh context per iteration using task sets and memories. Also supports status, init, taskset management, add-task, and memories subcommands.
/ralphOrchestrates tasks to completion by delegating to tiered subagents (architects, executors, designers) with parallel execution, model routing, and architect verification loop.
运行 oma ralph 模式。
目标： $ARGUMENTS
Ralph 协议：
team-plan -> team-prd -> team-exec -> team-verify -> team-fix。team-verify。team-fix 然后返回到 team-verify。模式姿态：
输出格式：
| 门控 | 状态 | 证据 |
|---|---|---|
| team-plan | ... | ... |
| team-prd | ... | ... |
| team-exec | ... | ... |
| team-verify | ... | ... |
| team-fix | ... | ... |