claude-code-skill

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). High-risk: the code intentionally runs untrusted MCP server packages with full environment propagation, exposes mechanisms to register arbitrary webhooks, and includes flags to bypass permission checks — collectively enabling credential exposure, remote code execution, and supply-chain/backdoor abuse.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill explicitly enables running arbitrary shell commands and file writes (Bash, Write), includes a dangerous "bypassPermissions" mode, and shows adding system dirs (e.g. /var/log) — together these permit modifying system files or bypassing prompts, so it can be used to compromise the host.