Skip to content
Agentmart
Submit a tool

herdr

Control Herdr panes, layouts, commands, and agents only when explicitly requested; requires HERDR_ENV=1.

Skill by lstmemery

IndexedMIT★ 1
Write a review

Chalk Labs

Need help deploying herdr?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/lstmemery-herdr

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Control Herdr panes, layouts, commands, and agents only when explicitly requested; requires HERDR_ENV=1.

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/herdr

What's inside

Skill name
herdr
Description
Control Herdr panes, layouts, commands, and agents only when explicitly requested; requires HERDR_ENV=1.
License
MIT
Instructions
38 lines in SKILL.md · view on GitHub

Files (5)

  • AGENT-CONTROL.md
  • COMMANDS.md
  • REMOTE.md
  • SKILL.md
  • trigger-cases.json · tests/

SKILL.md (first lines)

# Herdr

Confirm the requested state or result from live responses, report the target by name or label, and clean up only task-owned temporary layout or state why it remains. Agent lifecycle readiness alone is not proof the task is complete.

## Enter and discover

Before any control command, check `test "${HERDR_ENV:-}" = 1`. If it fails, report that this agent is outside Herdr and stop; do not control another client's focused session.

Use `herdr --help`, then the relevant bare **command group** (`herdr agent`, `herdr pane`, `herdr workspace`, `herdr tab`, `herdr worktree`, `herdr terminal`, `herdr notification`, `herdr integration`, `herdr session`, or `herdr machine`). Installed help owns syntax. Bare `herdr` launches or attaches the TUI; do not use it for discovery. An incomplete-looking mutating command such as `herdr workspace create` can execute with defaults; never probe mutations by omitting arguments.

Read IDs and state from returned JSON. Server errors are JSON on stderr with exit 1; syntax errors exit 2.

## Target and topology

- Use `--current` for the calling pane, an explicit pane ID, or a unique live agent name. Omitted targets can use another client's UI focus.
- Workspace (`w1`), tab (`w1:t1`), and pane (`w1:p1`) IDs are opaque; closed tab/pane IDs are not reused. Discover with `workspace list`, `tab list --workspace "$HERDR_WORKSPACE_ID"`, `pane current --current`, `pane list --workspace "$HERDR_WORKSPACE_ID"`, or `agent list`.
- **Moves change IDs.** After `pane move`, use `.result.move_result.pane.pane_id` or the live agent name. `.result.move_result.previous_pane_id` remains usable only through the moved process's inherited caller context, not as a general target.
- Creation returns the next handles: `workspace create` → `.result.workspace`, `.result.tab`, `.result.root_pane`; `tab create` → `.result.tab`, `.result.root_pane`; `pane split` → `.result.pane`.
- Use the topology and working directory explicitly requested for the task. When Orchestrator is the caller, apply its selected workspace/worktree policy. Otherwise default to a sibling pane in the current tab and current directory; do not invent a new workspace or worktree.
- Preserve focus with `--no-focus` for background work unless the user asked to switch. For a sibling split, honor the requested direction; otherwise inspect `herdr pane layout --pane "$HERDR_PANE_ID"`, split wide panes right and narrow/tall panes down, and avoid repeated splits that leave unusable geometry. Preserve cwd explicitly, for example `herdr pane split --current --direction right --cwd "$PWD" --no-focus`.

Use pane commands for raw terminals and ordinary processes; use agent commands for recognized agents and lifecycle-aware control. `agent start` needs an existing available shell pane and never creates or moves layout.

## Select the operation

- **Start, prompt, wait for, or interact with an agent:** read [AGENT-CONTROL.md](AGENT-CONTROL.md) before control input.
- **Run an ordinary command or interpret terminal output:** read [COMMANDS.md](COMMANDS.md). Agent-output recovery uses its read guidance too.
- **Remote hosts, saved machines, or multiple servers:** read [REMOTE.md](REMOTE.md) before selecting targets or changing profiles.

## Ownership and recovery

Inspect state and output before retrying a failed wait or prompt. A timeout or stalled response does **not** prove nondelivery; never blindly resubmit work.

Close only workspaces, tabs, panes, and sessions you created unless the user requested broader cleanup. `workspace close --group` also closes linked worktree workspaces; it is not a retry for `workspace_group_close_required`.
Route requests to clear stale panes through `herdr-jobs cleanup-stale` so managed leases are reconciled before workspaces close.
For a requested workspace close, use `herdr-jobs close-workspace WORKSPACE_ID`; it refuses held managed leases and names the run directory.

`--trust-repository` grants per-request Git trust; use it only after user verification of that repository. Check `herdr status` before relying on new server features after an update. A missing method does not authorize stopping or replacing the server. `herdr server stop` kills pane processes: require the user's explicit intent to stop them. Never kill the main Herdr process; isolated experiments use named test sessions.

From lstmemery/skills, MIT.

More from lstmemery/skills

Signals

Stars
1
Forks
0
Open issues
0
Last commit (this folder)
Oct 19, 2026 (11 days ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from lstmemery

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get herdr running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying herdr?

CallGet help
Cite this page

herdr skill. Agent Mart. https://agentmart.dev/skills/lstmemery-herdr (last updated 2026-10-24)