Claude Code allowed to run any shell command
HighAV-FW-012ToolsFramework packNew in 2.0
What it finds
Claude Code settings allow every Bash command without asking ("Bash" or "Bash(*)").
Why it matters
Any command the model decides to run — including ones suggested by files or web pages it reads — runs without your approval.
What an attacker could do: The agent can run any command on this machine without asking.
How to fix it
Allow only specific commands (for example "Bash(npm test)", "Bash(git status)") and leave everything else on ask.
Pro shows the corrected code for your exact file, step by step, and re-scans to confirm the fix. See plans
Check details
- Finding code
- CLAUDE_CODE_BROAD_BASH_ALLOW
- How it detects
- Looks for a specific risky pattern in the code or config.
- Confidence
- Definite: the risky setting is written plainly in the file.
- Scans
- Agent source using the named framework (LangChain, CrewAI, AutoGen, OpenAI Agents SDK, Claude Agent SDK, computer-use tools)
- OWASP LLM Top 10
- LLM08 - Excessive Agency
- NIST AI RMF
- MG-2.2
- SOC 2
- CC6.1, CC6.8
Check your agent for this
Paste your agent code or MCP config. The free scan runs all 124 checks in about 10 seconds — no account needed.