D65/100

codex-mcp-server
mcp1.4.2MCP server wrapper for OpenAI Codex CLI that enables Claude Code to leverage Codex's AI capabilities directly.
By tuannvm | 8 findings | Scanned 3/22/2026 | tooltrust-scanner/v0.1.15
1 Critical3 High2 Medium2 Low
Security Findings (8)
- CriticalArbitrary Code Execution
codex:tool name or description implies arbitrary script/code execution (evaluate_script, execute javascript, etc.)
Rule: AS-006Fix: This tool can execute arbitrary code or shell commands on the host system. Remove it unless strictly required. If kept: (1) restrict access to trusted users/agents only, (2) require human approval before each invocation (Claude Desktop: set approval_required: true; other clients: enable equivalent confirmation), (3) use the most restrictive sandbox or read-only mode available, and (4) never expose this tool to untrusted input sources. - HighExcessive Permission Surface ×3
- codex:tool declares network permission
- codex:tool declares exec permission
- websearch:tool declares network permission
Rule: AS-002Fix: Tool requests broad permissions (exec/fs/network). Validate input parameters using Enums where possible, and restrict file system operations to explicit allowed directories. - MediumExcessive Permission Surface ×2
- codex:tool declares fs permission
- review:tool declares fs permission
Rule: AS-002Fix: Tool requests broad permissions (exec/fs/network). Validate input parameters using Enums where possible, and restrict file system operations to explicit allowed directories. - LowDoS Resilience — Missing Rate Limit / Timeout ×2
- codex:tool performs network or execution operations but declares no rate-limit, timeout, or retry configuration
- websearch:tool performs network or execution operations but declares no rate-limit, timeout, or retry configuration
Rule: AS-011Fix: Declare explicit rate-limit, timeout, and retry configuration for all network and execution tools. Implement exponential back-off and surface resource state to the calling agent.
Scan this tool yourself
Reproduce this audit locally or block risky tools in CI.
Install once, then scan any MCP server:
$ curl -sfL https://raw.githubusercontent.com/AgentSafe-AI/tooltrust-scanner/main/install.sh | bash$ tooltrust-scanner scan --server "npx -y codex-mcp-server"Adjust the package name if your npm registry name differs from the tool ID. View source
Add badge to your README
Copy this Markdown to show your ToolTrust grade on GitHub.
[](https://github.com/AgentSafe-AI/tooltrust-directory)