core

Name: wsh/core
Author: deepgram

Share on X Share on LinkedIn Share on Bluesky

40Medium

This skill performs unauthorized network exfiltration via curl and wget while bypassing security constraints by failing to declare its tool usage or network capabilities.

ThreatsCI SM CS BC S HITL

Threat Analysis

AI Agent Traps ↗Scanned 6/20/2026

Content InjectionPerceptionclean

Semantic ManipulationReasoningclean

Cognitive StateMemory & Learningclean

Behavioural ControlAction1 finding

SystemicMulti-Agent Dynamicsclean

Human-in-the-LoopHuman Overseerclean

Skill Info

Namedeepgram/wsh/core

Registrygithub

Version4863aafbf070

PURLpkg:github/deepgram/wsh@4863aafbf070?skill=core

Stars5

Installs1

Source

Repository ↗SKILL.md ↗

Install

This skill has findings worth reviewing before installing.

GitHub

1 installs

Skills.shDocs

npx skills add https://github.com/deepgram/wsh

Assessments (3)

AI Agent Traps ↗

Pipe to curl/wget for data exfiltration detectedmediumdata exfiltrationAML.T0025AML.T0024LLM02:2025

Pipe to curl/wget for data exfiltration detected (seen 7 times in this file at lines 80, 126, 127, 128, 129, 130, 387)

100% confidenceline 80

curl -s -X POST --unix-socket $WSH_SOCK http://localhost/sessions/work/input -d $

Missing allowed-tools declarationlowunauthorized tool use

Skill executes commands, writes files, or accesses the network but declares no allowed-tools, so its tool surface cannot be reviewed or constrained.

90% confidence

Undeclared network usagelowunauthorized tool use

Skill contains network access patterns (HTTP requests, socket connections) without declaring network capability in allowed-tools.

60% confidence

Missing usage exampleslowpolicy violation

Skill body contains no code blocks or usage examples, making it harder for users to evaluate.

100% confidence

Missing licenseinfopolicy violation

Skill does not specify a license field. Specifying a license helps users understand usage terms.

100% confidence

Missing or insufficient descriptioninfopolicy violation

Skill description is empty or too short. A clear description helps users evaluate the skill's purpose.

100% confidence

Badge

Markdown

[![Mondoo Skill Check](https://mondoo.com/ai-agent-security/api/badge/github/deepgram/wsh/core.svg)](https://mondoo.com/ai-agent-security/skills/github/deepgram/wsh/core)

HTML

<a href="https://mondoo.com/ai-agent-security/skills/github/deepgram/wsh/core"><img src="https://mondoo.com/ai-agent-security/api/badge/github/deepgram/wsh/core.svg" alt="Mondoo Skill Check" /></a>

Image URL

https://mondoo.com/ai-agent-security/api/badge/github/deepgram/wsh/core.svg

Secure your AI agents

Skills can read files, run commands, and access credentials. Mondoo helps organizations manage the security risks of AI agent skills across their entire fleet.

Continuous scanning of skills across all registries
Policy enforcement before skills reach your agents
Integration with your existing security workflow