TermherdTermherd/termherd/.claude/skills/run-termherd/SKILL.md

run-termherd

Skill2 starsChanged 6 days agoMITView on GitHub

Launch a termherd build to see a change working, beside the user's own running termherd, and observe it through tracing, a capture or the MCP bridge. Use when asked to run, start or screenshot termherd, to check a GUI or OS-level behaviour no headless test reaches (quit flow, window, canvas), or to test a worktree build.

In this repository5 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- drive-termherdSkill
- file-issueSkill
- mutation-testSkill

The file

Create a free account to mount

Files

SKILL.md

Run termherd

The full reference is docs/agents/running.md; read it before the first launch in a session. The steps that matter:

The user's release termherd usually hosts this session and holds the single-instance lock. Isolate the build's temp dir and home, or it refuses to start or repaints the user's app with its settings:
```
USERPROFILE=$(mktemp -d) TMPDIR=$(mktemp -d) RUST_LOG=info cargo run -p termherd-app
```
On Windows set TMP to a directory that exists, never a missing one. Copy ~/.claude/projects into the fake home when the check needs real sessions.
Launch it in the background and read its tracing output; add info! at the seam you are verifying, never println!.
A build an agent launches inherits this session's CLAUDE* variables, so a Claude started inside it writes no session file. Anything reading Claude's session files is verified from a build the user starts.
A branch made before the lock moved does not see a newer build: build the merged tree (git merge --no-commit origin/main, build, git merge --abort).
To observe: ⌘⇧S / Ctrl+Shift+S writes a JSON + PNG capture to ~/.termherd/captures/ (the latest is the highest name); the MCP bridge reads a live app — see the drive-termherd skill.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free