How do I install the Runboard skill?

Run npx skillmds@latest add whamp/runboard in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Runboard safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Runboard?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Runboard free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Runboard?

whamp (@whamp) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Runboard by whamp — Agent Skill

Name: Runboard
Author: whamp

Runboard

Open a Herdr tail tab for a harness run when the user explicitly asks for a runboard, tail, Herdr tab, or raw log view. For new harness/run_batch.py monitoring, prefer the structured dashboard at scripts/run_dashboard.py reading results/_runs/<run_id>; this skill is the compatibility/tail workflow.

whamp 708cd75 2 files · 4.2 KB Updated Sep 17, 2026

File contents

runboard

A runboard is a live .out file plus a Herdr tab tailing it. For new batch launches, prefer the structured dashboard:

python3 scripts/run_dashboard.py --state-root results/_runs --port 8765

Use this skill when the user specifically wants a Herdr/tail view or when a legacy run only has runs/<run>/track.out. Prefer the helper; do not hand-roll Herdr discovery unless the helper fails.

Native line shape:

[71/113] onedump-dump-encryption-pipeline / observational-memory / rep0  ok

Status legend: ok | empty | timeout | transient | exit=<n>.

Tight path: open an existing live track

Use this path when runs/<run>/track.out already exists, especially for legacy harness/run_batch.py ... | tee runs/<run>/track.out launches that predate structured state.

Run the helper:

scripts/open_runboard.py --run '<run>' --label '<short label>'
# or: scripts/open_runboard.py --track runs/<run>/track.out --label '<short label>'

Report only the returned pane, track, reused, and current progress from visible_preview.

Completion: helper output has "status": "live". The helper has already proved Herdr env, current workspace, exact tail -f process, and visible runboard content.

If there is no track.out

Only for historical/result-only runs, build a compatibility tracker first. Read result-tracker.md, then rerun the tight path.

Completion: runs/<run>/track.out exists and the helper returns status=live.

Guardrails

If the helper says HERDR_ENV != 1, stop; do not control Herdr from outside Herdr.
Do not use runboard as the default monitor for a new run if structured results/_runs/<run_id>/ state is available; use the dashboard instead.
Do not dump panes, containers, token counts, streamed text, or repeated state snapshots. A runboard is one native line per cell.
Do not claim success from a pidfile or existing track.out; success requires the helper's verified status=live result.
For single harness/run.py cells, there is no batch runboard; tail that cell's logs only if asked.

---name: runboarddescription: Open a Herdr tail tab for a harness run when the user explicitly asks for a runboard, tail, Herdr tab, or raw log view. For new harness/run_batch.py monitoring, prefer the structured dashboard at scripts/run_dashboard.py reading results/_runs/<run_id>; this skill is the compatibility/tail workflow.---# runboardA **runboard** is a live `.out` file plus a Herdr tab tailing it. For new batchlaunches, prefer the structured dashboard:```shpython3 scripts/run_dashboard.py --state-root results/_runs --port 8765```Use this skill when the user specifically wants a Herdr/tail view or when alegacy run only has `runs/<run>/track.out`. Prefer the helper; do not hand-rollHerdr discovery unless the helper fails.Native line shape:```txt[71/113] onedump-dump-encryption-pipeline / observational-memory / rep0 ok```Status legend: `ok` | `empty` | `timeout` | `transient` | `exit=<n>`.## Tight path: open an existing live trackUse this path when `runs/<run>/track.out` already exists, especially for legacy`harness/run_batch.py ... | tee runs/<run>/track.out` launches that predatestructured state.1. Run the helper: ```sh scripts/open_runboard.py --run '<run>' --label '<short label>' # or: scripts/open_runboard.py --track runs/<run>/track.out --label '<short label>' ```2. Report only the returned `pane`, `track`, `reused`, and current progress from `visible_preview`.Completion: helper output has `"status": "live"`. The helper has already provedHerdr env, current workspace, exact `tail -f` process, and visible runboardcontent.## If there is no track.outOnly for historical/result-only runs, build a compatibility tracker first. Read[`result-tracker.md`](result-tracker.md), then rerun the tight path.Completion: `runs/<run>/track.out` exists and the helper returns `status=live`.## Guardrails- If the helper says `HERDR_ENV != 1`, stop; do not control Herdr from outside Herdr.- Do not use runboard as the default monitor for a new run if structured `results/_runs/<run_id>/` state is available; use the dashboard instead.- Do not dump panes, containers, token counts, streamed text, or repeated state snapshots. A runboard is one native line per cell.- Do not claim success from a pidfile or existing `track.out`; success requires the helper's verified `status=live` result.- For single `harness/run.py` cells, there is no batch runboard; tail that cell's logs only if asked.

Frequently asked questions