herdr

Configure or tune herdr — keybindings, theme, agent-detection rules, notifications, or write a herdr plugin.

Save, restore, or design a reusable herdr pane/tab layout.

Reattach to herdr or recover a session after a restart or update — explain what survived.

Triage your herdr agent fleet — find which agents are blocked, done, or need attention.

Set up a herdr workspace and herd a fleet of agents to work a multi-part task in parallel.

Read-only triage of a herdr agent fleet. Use to find which agents are blocked, done-but-unreviewed, or stuck, and to explain why a herdr agent's detected state looks wrong. Inspects and reports only — never sends input, spawns, or restarts agents.

Sets up a herdr workspace and dispatches a fleet of coding agents to work a multi-part task in parallel, then coordinates their handoffs. Use when a task should be split across several agents in herdr. Spawns and assigns agents (without stealing focus) and wires up wait-coordination between them.

Monitors and triages a fleet of agents in herdr, and explains why a herdr agent's detected state looks wrong. Teaches the detection-AUTHORITY model (lifecycle-state hooks vs screen-manifest detection), state roll-up across pane/tab/workspace, a triage workflow for finding who needs attention, and a debugging path with `herdr agent explain`. Defers raw command syntax to the vendored herdr skill and live `herdr --help`. Triggers: 'which herdr agent needs attention', 'why does herdr show my agent as idle/blocked', 'monitor my agents status', 'an agent is stuck/blocked in herdr', 'herdr agent state is wrong / explain detection', 'is my agent done in herdr'.

Control herdr from inside it. Manage workspaces and tabs, split panes, spawn agents, read output, and wait for state changes — all via CLI commands that talk to the running herdr instance over a local unix socket. Use when running inside herdr (HERDR_ENV=1).

Use when configuring or tuning herdr through its config file or detection manifests — editing ~/.config/herdr/config.toml, changing keybindings (prefix mode, direct shortcuts, special keys, custom pane/shell/plugin_action commands), setting a theme or custom colors, tuning agent-detection/labeling rules, configuring notifications and sounds, terminal/shell and sidebar/scrollback/mouse behavior, experimental flags, or authoring herdr's own plugins (herdr-plugin.toml). Triggers: 'configure herdr', 'change herdr keybindings', 'rebind herdr prefix', 'set a herdr theme', 'tune herdr agent detection rules', 'herdr notifications/sounds', 'edit config.toml for herdr', 'write a herdr plugin', 'create herdr-plugin.toml'. Does NOT cover in-pane control CLI, session/workspace lifecycle, or the meaning of agent states.

Use when designing, saving, or reproducing a pane/tab arrangement in herdr — 'save/restore my herdr layout', 'set up a reusable pane layout in herdr', 'recreate this workspace layout', 'export a herdr layout tree', 'declarative workspace setup in herdr', or capturing an editor+server+logs+tests arrangement as a shareable, reproducible artifact. Also covers the judgment around resize/swap/move/zoom for legibility. Use this skill to arrange, save, or reproduce panes *now*; for session lifecycle, recovery, and which-container-to-use judgment use herdr-workspace-management. Defers raw command syntax to the vendored herdr skill and the live socket API; defers keybindings to herdr-configuration.

Use when herding a fleet of coding agents in herdr — deciding whether to run several agents in parallel vs a single agent, setting up an agent council/team, having one agent coordinate or manage others, routing input to a specific agent by name or pane, dividing work so agents don't collide, making agent B wait on agent A, spawning helper agents without stealing focus, or running a fleet on a remote box. Holds the durable judgment of multi-agent assignment; defers all raw command syntax to the vendored herdr skill and the live CLI. Triggers: 'run several agents in parallel in herdr', 'set up an agent council/team', 'have one agent coordinate others', 'herd agents on a remote box', 'split a task across multiple coding agents', 'route this to a specific agent', 'make one agent wait for another'.

1 hook across 1 event