Back to skill

Security audit

Alfred OpenShell Sandbox

Security checks for vulnerabilities and agentic risk

Overview

This skill is a straightforward OpenShell sandbox usage guide with some operational risks users should understand, but no hidden or purpose-mismatched behavior was found.

Before installing, users should be comfortable with OpenShell sandbox administration. Confirm sandbox names before deletion, treat copied files from sandboxes as untrusted until reviewed, and only loosen network policies when necessary for the task.

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
  • YARA SignaturesMalware Match, Webshell Match, Cryptominer Match
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

YARA rule 'offensive_tool_references': References to well-known offensive security tools [hacktools]

High
Category
YARA Match
Confidence
70% confidence
Finding

YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).

Content

Scanner excerpt · SKILL.md (reported line 61)May include surrounding context.

Integration

For Coder Agent

When executing code that could affect the host system:

bash
# Instead of running locally:
python3 script.py

# Run in sandbox:
openshell sandbox exec -n coder-sandbox -- python3 /workspace/script.py

For Security Agent

When running security tools or scans:

bash
# Run nmap, nikto, etc. in isolated sandbox
openshell sandbox exec -n security-sandbox -- nmap -sV target

For Debug Agent

When reproducing bugs or testing fixes:

bash
openshell sandbox exec -n debug-sandbox -- node test.js

For QA-Tester

When running test suites:

bash
openshell sandbox exec -n test-sandbox -- pytest tests/

File Transfer

To copy files between host and sandbox:

bash
# Copy file INTO sandbox (via exec cat)
cat local_file.py | openshell sandbox exec -n coder-sandbox -- tee /workspace/local_file.py

# Copy file FROM sandbox
openshell sandbox exec -n coder-sandbox -- cat /workspace/result.txt > local_result.txt

Policies

Default

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The markdown documents openshell sandbox delete <name> as a CLI action, but does not warn that deleting a sandbox is irreversible and will remove its ephemeral disk contents. Because markdown files should disclose behaviors that may affect user data or system integrity, this omission is a safety-relevant missing warning.

Content

No source excerpt is available for this finding.