claude-codex-dispatch

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly instructs the agent to "surface the target's raw recovery handle as soon as it exists," which requires outputting a sensitive handle/token verbatim and therefore presents a direct secret-exfiltration risk.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill includes a runtime installer command that fetches and pipes a remote script to the shell (curl -fsSL https://claude.ai/install.sh | bash), which would execute remote code and is presented as the way to install a required CLI dependency.