Security audit

AgentMesh Governance

Security checks for vulnerabilities and agentic risk

Overview

The skill has a coherent governance purpose, but its shell wrappers handle user input unsafely enough to warrant manual review before installation.

Review this skill before installing. Only run it in a constrained environment, avoid passing untrusted agent names, policy paths, messages, signatures, or action strings to the scripts, and prefer a version that passes inputs through argv or stdin, validates arguments, pins dependencies, and fails closed when AgentMesh is unavailable.

A.I.G

Vulnerability Patterns

Insecure DependenciesIntroduces malicious components through unsafe dependency sources
Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review

Findings (2)

T09 · Insecure Skill Coding Practices

Error

Location: scripts/audit-log.sh:15
Finding: Arbitrary Python Code Execution Through Unsafe Argument Interpolation
Content: View full analysis
Remediation: View remediation

T08 · Insecure Dependencies

Warning

Location: SKILL.md:30
Finding: Unpinned Third-Party Package and Mutable Git Dependency Installation
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 95% confidence
Finding: The activation criteria are overly broad because they trigger on general questions about agent safety, governance, compliance, or trust rather than narrowly scoped operational tasks. In an agent runtime, this can cause the skill to activate in unrelated conversations, increasing the chance it steers decisions, invokes scripts, or inserts policy logic where the user did not explicitly request this governance layer.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 99% confidence
Finding: The shell script interpolates the user-controlled AGENT value directly into a Python -c string enclosed in double quotes and then into a Python single-quoted literal. An attacker can supply crafted input containing quotes and Python syntax to break out of the string and achieve arbitrary Python code execution when the script runs.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The fallback path returns a hardcoded trust score of 0.5 and labels it as baseline trust information, which can mislead downstream automation or operators into treating an unverified agent as moderately trusted. In a governance skill, trust scores may drive delegation, policy bypass, or access decisions, so placeholder data can directly undermine security controls even without code execution.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.