Orchestration

Oversee parallel agent threads on a project, merge finished work and restart stalled ones

Submit ideas, pick up work and deliver paid services on the OpenServ Ideaboard

Orchestrate multi-agent workflows with task dependencies and shared workspaces

Run an autonomous loop that modifies, verifies and keeps or discards changes

Route each request to the narrowest matching tool family and its verification step

Turn multi-file rewrites and migrations into a spec and verify the result

Cancel stuck or orphaned Ouroboros executions

Start or monitor an evolutionary loop that improves a codebase over time

Execute a Seed spec and verify every acceptance criterion it defines

Check Ouroboros session status and how far work has drifted from the goal

Run a background loop of evolve steps that keeps working until done

List in-flight Ouroboros sessions and re-attach after an MCP disconnect

Run an autonomous loop of experiments that keeps going until a metric improves

Generate several candidate solutions, score them on a frozen rubric and keep the top one

Audit Claude Code and Codex subagent files for loadability, scope and overlap

Run a full feature, refactor or hard bug through the CCG analyze, design and verify roles

Design multi-agent teams with defined roles, lifecycles and task decomposition

Coordinate multiple agents with task decomposition, parallel runs and conflict resolution

Delegate coding tasks to coding agents and follow up until code is tested

Group delegated work into a task graph and dispatch it across roster agents

Track a job split across detached tasks as one flow with a single owner

Triage inbox messages into immediate alerts, waiting replies and later summaries

Build a repo-local skill and a GitHub Actions workflow for iterated coding agents

Design a control loop for your codebase and build it as a scheduled coding agent

Replay or fork a past agent run from its recording to explain or reproduce it

Delegate focused coding, research or review work to a bounded subagent

Route plain-language requests to Codex, Cursor, Gemini CLI and other coding sessions

Run OpenProse .prose files that orchestrate multi-step workflows across agents

Pick the right agent or crew for a task and delegate parallel work to it

Have ChatGPT design and review a local project while Codex writes and tests the code

Get a second opinion or web search from xAI Grok without switching your main model

Run long agent tasks in background jobs so chat stays responsive

Add, list and remove scheduled cron jobs for agent workflows in claw

Delegate recurring checks and follow-ups to the heartbeat agent via HEARTBEAT.md

Run prompts on a set interval and get notified of the results

Convene a council of historical thinker personas to analyze a hard problem

List testable code surfaces in focus files as candidates for automated fixes

Break a large goal into work items, dispatch sessions and verify each claim

Run a goal-driven loop that keeps working until the definition of done is met

Convene a council of models from different vendors to vote on hard questions

Manage an issue and pull request pipeline on one repo by sending intents to workers

Check your reasoning by explaining the work to a fresh reviewer who finds gaps

Keep one long-running session working toward a goal with scheduled self-nudges

Run a goal as a team by deciding what to do yourself and dispatching the rest

Monitor and audit another coding agent's work from its session, PR or log

Delegate research, coding and tests to cheaper subagents to save cost

Keep planning with an expensive model and delegate bounded work to cheaper subagents

Compare and merge competing plans from several coding agents into one

Pause parallel agent runs near usage limits and resume when they reset