Psyseal scan

io.github.psysealPublished by GitHub account psysealv0.3.3Updated Nov 11, 2026

Offline check of your AI agent: heartbeat cost, plaintext keys, planted or bloated memory.

The publisher's own server: scan.mx Email Diagnostics

VerifiedSTDIODesktop onlySecurity & MonitoringAI & ML

Overview

AI-generated overview

A local, read-only MCP server that scans AI agent folders on your machine for exposed plaintext secrets, costly heartbeats and bloated or planted memory.

What you can do
  • Run one read-only scan tool, scan_agents, over stdio against the AI agents installed on the computer.
  • Report what loads on every message, what heartbeats and scheduled tasks cost, and which plaintext secrets sit in memory or skills.
  • Flag memory that looks planted or bloated, with a concrete fix for each finding.
  • Show the type, file and line of each secret without revealing its value or quoting memory text.
When to use it
Useful when you run one or more AI agents locally (such as OpenClaw, Hermes Agent, nanobot, PicoClaw, ZeroClaw or NanoClaw) and want a quick offline audit of secret exposure, per-message context cost and memory hygiene before sharing or deploying an agent setup.
Requirements
Runs as a local stdio process via npx, so Node.js 20 or newer is needed. No network access, dependencies, accounts, API keys or environment variables are required. A folder path can be passed to check any agent folder.
Before you install
The scan reads agent files on your computer, so it sees whatever those folders contain, including secrets; the report is designed to show only type, file and line, never values. It is described as read-only, with no network use, no file writes or deletions and no symlink following. The free scanner's report ends with one promotional line about the paid Psyseal package, which can be turned off.

Installation

In SourceWeft

  1. Open Psyseal scan in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Desktop only via STDIO. STDIO servers start a local process, so they need the SourceWeft desktop host.

Other MCP clients

Follow the launch instructions in the repository.

README

psyseal-scan

The free Psyseal scanner. MIT-0.

It reviews the AI agents on your computer (OpenClaw, Hermes Agent, nanobot, PicoClaw, ZeroClaw, NanoClaw or any agent folder) and tells you what loads on every message, what heartbeats and scheduled tasks cost, which plaintext secrets are sitting in memory or skills, and which memory looks planted or bloated. Every finding comes with a concrete fix.

It runs on your computer, reads files only, uses no network and has no dependencies. Needs Node.js 20 or newer.

Run it

bash
npx psyseal-scan --lang en

Or from a copy of this repository: node scripts/psyseal-scan.mjs --lang en. See --help for all options (--path <folder> checks any agent folder).

As an agent skill

  • OpenClaw (from ClawHub): openclaw skills install @psyseal/psyseal-scan
  • Claude Code, Cursor, Codex, Gemini CLI and other agents (skills CLI): npx skills add psyseal/psyseal-scan
  • Claude Code plugin (the skill): /plugin install psyseal-scan --marketplace psyseal/psyseal-scan (before Claude Code 2.1.275: /plugin marketplace add psyseal/psyseal-scan, then /plugin install psyseal-scan@psyseal)

As an MCP server

The same scan as one read-only tool, scan_agents, over stdio:

json
{  "mcpServers": {    "psyseal-scan": { "command": "npx", "args": ["-y", "psyseal-scan", "mcp"] }  }}

Claude Code: claude mcp add psyseal-scan -- npx -y psyseal-scan mcp

Name in the official MCP Registry: io.github.psyseal/psyseal-scan.

What it never does

It never sends anything over the network, writes or deletes files, starts other programs or follows symlinks. The report shows the type, file and line of each secret, never its value, and never quotes memory text.

This repository holds only the free scanner. The full Psyseal package, including the paid cleanup and the signed seal, is at https://github.com/psyseal/scanner, and the service at https://psyseal.com. The report ends with one line about it; --no-cta removes it.


Español: el escáner gratis de Psyseal (MIT-0). Revisa en tu computadora cuánto te cuestan los latidos y tareas programadas de tu agente, qué claves quedaron expuestas y qué memoria está inflada o trae instrucciones plantadas. No usa la red ni tiene dependencias. Córrelo con npx psyseal-scan --lang es. También funciona como skill, como servidor MCP (npx -y psyseal-scan mcp) y como plugin de Claude Code (/plugin install psyseal-scan --marketplace psyseal/psyseal-scan). Más información en https://psyseal.com.

Source: README.md at commit ecc8601

Tools

0
SourceWeft has not started this package in its sandbox yet, so its tools are not known.

Version history

1
  1. v0.3.3LatestNov 11, 2026

Score

50% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 days25 pts

    Keep the endpoint answering the MCP handshake; SourceWeft probes it every 30 days.

  • Auth declared10 pts

    Declare the credential headers in the registry entry, or keep the endpoint answering the probe so its auth can be classified.

  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts

Other Scan integrations