ClawHub
Back to skill

Security audit

Work over SSH

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed SSH workflow for working on a user-specified remote Git project and does not show hidden persistence, credential theft, or deceptive behavior.

Install this only if you want an agent to use your existing SSH access to inspect, test, and patch a remote Git repository. Provide only non-secret connection details, review requested changes, and require explicit approval for package installs, deployments, service restarts, database changes, or destructive Git/filesystem actions.

A.I.G

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
82% confidence
Finding
The skill clearly instructs use of local shell and file-reading capabilities (`ssh`, `python3`, helper scripts, patch application), but it does not declare an explicit tool scope such as `permissions` or `allowed-tools`. In agents that rely on manifest-level scoping, this can cause the runtime to grant broader-than-intended access or leave reviewers without an enforceable boundary, increasing the risk of command execution beyond the minimum needed for the skill.

subprocess module call

Medium
Category
Dangerous Code Execution
Content
capture: bool = False,
) -> subprocess.CompletedProcess[bytes]:
    command = [*SSH_BASE, host, remote_command]
    return subprocess.run(
        command,
        input=stdin,
        check=False,
Confidence
70% confidence
Finding
subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Static analysis

No suspicious patterns detected.