agentleFS
NecturaLabsNecturaLabs/AgentSkills/skills/open-project-terminal/SKILL.md

open-project-terminal

Skill1 starsChanged 51 days agoMITView on GitHub

Open a new Claude Code session in its own terminal window for a named project, when the user asks to open up a session, terminal or Claude for a project (for example "open a terminal for Canopy"). Launches the terminal on a chosen workspace, in the project's directory, with the project's name and a first message. Not for opening files or an editor, not for running commands in the current session, and not for starting subagents (agent-orchestration).

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When the user names a project that has no folder yet and says it is new, create the empty folder under the projects root (never git init …

    line 24

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Open a project session terminal

The defaults below are this setup's; a value the user states overrides its default, and you change the projects root, terminal and workspace once for your machine.

SettingDefault
Terminalkitty, window title = project name
Directorythe project's path under your projects root, <projects-root>/<Name> (find it with ls <projects-root>; if no folder matches, say so and stop)
Permissions--dangerously-skip-permissions (bypass)
Model, effort--model haiku --effort low
Session nameproject name, --name <Name>
First messagehi, passed as the prompt argument
Workspace1, opened silently so focus does not move

A new project directory

When the user names a project that has no folder yet and says it is new, create the empty folder under the projects root (never git init unasked). A new directory makes the session stop at Claude Code's folder-trust prompt, so the first message never arrives and the session stays out of the app's recents. Trust it before launching, so the user never has to tend a prompt:

  1. Copy ~/.claude.json to a scratch directory as a backup.
  2. Set projects["<absolute directory>"].hasTrustDialogAccepted = true in it: load the JSON, write a temporary file in the same directory with the original file mode, then os.replace it over the original.

Skip this for a directory that already has sessions. Never answer the prompt by sending a key to the window: a dialog's default can be "exit" and closes the session.

Command

Recent Hyprland takes Lua dispatch (the old exec [workspace 1 silent] ... syntax errors):

hyprctl dispatch 'hl.dsp.exec_cmd("kitty --title <Name> --directory <path> claude --dangerously-skip-permissions --model haiku --effort low --name <Name> hi", {workspace="1 silent"})'

Kill a window you must close by its exact PID, never pkill -f with a pattern that also appears in your own command line, which kills your shell. Then confirm the window exists on the workspace:

hyprctl clients -j | python3 -c "import json,sys;[print(c['class'],c['workspace']['id'],c['title']) for c in json.load(sys.stdin) if c['class']=='kitty']"

For a new directory also confirm ~/.claude/projects/<path with / as ->/*.jsonl exists, which means the first message went through. Report the project, path and workspace in one line. Do not touch other windows.

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.