pyx-scan
by fysoul17
Search & ResearchBrief
Check whether an AI agent skill is safe before installing or using it.
Description
What This Skill Does
Queries the PYX Scanner API to check whether an AI agent skill is safe before installing or using it. Returns a structured verdict, trust score, risk score, and recommendation based on the scan result. Handles safe, outdated, caution, failed, and unscanned states.
Automates safety verification in a single API call instead of manually reviewing skill source code or trusting the publisher's word.
When to Use It
- Checking a third-party skill before adding it to an agent
- Verifying a skill hasn't changed since its last safe scan
- Screening skills from unknown publishers before deployment
- Getting a risk score before using a skill in a production agent
- Confirming safety after a skill maintainer pushes an update
Example Workflow
User asks: Is anthropic/web-search safe to use?
- 1
Parse 'anthropic/web-search' into owner and name
- 2
Call the PYX Scanner API at /api/v1/check/anthropic/web-search
- 3
Read recommendation, trust score, risk score, and is_outdated fields
- 4
Determine verdict based on response conditions
- 5
Format and return a structured markdown safety report
A verdict of SAFE or CAUTION with trust score, risk score, confidence percentage, intent, and a summary of the skill's permissions and purpose
Installation
clawhub install fysoul17/pyx-scanRun the CLI command, or paste the prompt into Claude Code, Cursor, Codex or whichever coding agent you use.