usage: hirearchy [-h] [--root ROOT]
                 {init,hire,task,services,define-service,request,feedback,retire-service,plan,apply,send,status,health,progress,upgrade-rules,install-skill,review,inbox,resume,run,mcp,dash,obs,check,kiro-profile}
                 ...

A file-based organization of agents, run over MCP.

positional arguments:
  {init,hire,task,services,define-service,request,feedback,retire-service,plan,apply,send,status,health,progress,upgrade-rules,install-skill,review,inbox,resume,run,mcp,dash,obs,check,kiro-profile}
    init                create a new organization
    hire                create an agent (as root)
    task                give a task to an agent (as root)
    services            list the services you can request from your direct reports, with their
                        record
    define-service      define a standing service of a top-level agent, or amend its contract (as
                        root): on an existing service pass only what changes
    request             ask a top-level agent for one of its services (as root)
    feedback            accept or reject a delivered service request (as root)
    retire-service      take a service out of use (its files are kept)
    plan                validate a blueprint and print the plan for approval (exit 1 on errors)
    apply               create the agents of a blueprint and give its first task (as root)
    send                send a note to an agent (as root)
    status              show the organization tree
    health              show stuck, expensive, low-quality and idle agents and branches
    progress            one line per task you gave: state, who is busy, cost, last activity; what
                        needs you
    upgrade-rules       rewrite the rules in ORG.md to the current template (keeps settings and
                        mission)
    install-skill       copy the hirearchy assistant skill into a kiro skills directory (default
                        ~/.kiro/skills)
    review              run a health review now: tell managers (and you) about flagged agents
    inbox               read your (root) messages
    resume              unpause an agent
    run                 run the scheduler
    mcp                 run the MCP server on stdio (identity from $ORG_AGENT, default root)
    dash                terminal dashboard: agents, workflows, services, events, health (read-
                        only)
    obs                 read-only views of how the organization runs: now, trace, metrics, events
    check               invariants of the runtime: messages intact, budget conserved, nothing
                        stuck (exit 1 on FAIL)
    kiro-profile        write this organization's internal kiro profile (the scheduler does this
                        itself)

options:
  -h, --help            show this help message and exit
  --root ROOT           organization directory (default: $ORG_ROOT or nearest dir with ORG.md)
