Back to skill

Security audit

RedHat Command Execution

Security checks for vulnerabilities and agentic risk

Overview

This skill is a safety wrapper for terminal command execution, with disclosed risk checks and confirmation guidance rather than hidden code or automatic execution.

Review the malformed trigger metadata before publishing or installing so the skill only activates when intended. The command-running capability is broad by nature, so users should keep confirmations enabled for destructive, installer, network, and system-level commands.

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
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · SKILL.md (reported line 35)May include surrounding context.

md
2. Block/require confirmation for high-risk patterns:
   - recursive deletes (`rm -rf`, `Remove-Item -Recurse -Force`)
   - format/disk ops
   - piping remote scripts (`curl ... | sh`)
3. If supported, prefer dry-run flags first (e.g., `--dry-run`, `-n`, `--check`).
4. Execute the command and capture:
   - exit code

External Script Fetching

High
Category
Supply Chain
Confidence
90% confidence
Finding

Remote code is downloaded and executed. This bypasses code review and could introduce malicious code.

Content

Scanner excerpt · command-execution.md (reported line 25)May include surrounding context.

md
2. Block/require confirmation for high-risk patterns:
   - recursive deletes (`rm -rf`, `Remove-Item -Recurse -Force`)
   - format/disk ops
   - piping remote scripts (`curl ... | sh`)
3. If supported, prefer dry-run flags first (e.g., `--dry-run`, `-n`, `--check`).
4. Execute the command and capture:
   - exit code

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger definition is effectively unspecified/invalid (System.Object[]), which makes skill activation behavior ambiguous and difficult to constrain. In a command-execution skill, weak or malformed trigger constraints increase the chance the skill is invoked in unintended contexts, potentially exposing command-running capability without clear user intent or proper gating.

Content

No source excerpt is available for this finding.