Skip to content
Back to skills

Setup

ASecurity

Install or repair Persnally on this machine — the model of the user that this and every other AI client reads at session start. Run only when the user invokes /persnally:setup; it installs software and reads their local exports, sessions and repositories.

  • 6 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added September 23, 2026
ai-agentsgit

Works with

  • claude code
  • cursor
  • cli
  • mcp

Security analysis

A92/100
  • mediumInstalls packages at runtime which could introduce malicious dependencies

Pro shows the line behind each finding and how to fix it

Scanned September 23, 2026

npx -y skills add persnally/persnally --skill setup --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Setup?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Setup
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/persnally-setup/badge)](https://www.skillsdirectory.com/skills/persnally-setup)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
description: Install or repair Persnally on this machine — the model of the user that this and every other AI client reads at session start. Run only when the user invokes /persnally:setup; it installs software and reads their local exports, sessions and repositories.
disable-model-invocation: true
---

Persnally builds a model of the user from their own AI history, on their machine, and every
AI they use reads it. This skill gets it running here, non-interactively, and verifies it.

1. Check whether it is installed: run `command -v persnally && persnally --version`.
2. If it is not installed, run `npm i -g persnally`.
3. Run `persnally setup --yes`. This finds Claude/ChatGPT exports in `~/Downloads`, reads the
   local Claude Code, Cursor and Codex sessions and git repos, synthesizes the model, and
   connects the installed AI clients — including this one's MCP server, with its own identity
   token, which is why the tools are not bundled in the plugin. It picks a local Ollama model
   when one is available, so it runs fully offline; otherwise it asks for an Anthropic key —
   stop and ask the user for it rather than guessing. Never paste a key into chat. `setup`
   sees this plugin and does not add a second SessionStart hook.
4. Run `persnally doctor`. Report exactly what it prints. Exit code 1 means something is
   broken; say what, do not claim success.
5. Tell the user, in two lines, what Persnally learned about them (`persnally show profile`,
   first section only) and that this client now reads it at every session start via the
   plugin hook. Remind them: `persnally forget <topic>` deletes a claim and everything derived
   from it; `persnally export` takes everything with them.

If the user ran `persnally connect claude-code` on an older version before installing this
plugin, the SessionStart hook exists twice (user settings + plugin). Tell them to remove the
Persnally entry under `hooks.SessionStart` in `~/.claude/settings.json`; leave `~/.claude.json`
alone — that MCP entry, with its token, is the one the tools run through.

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…