api-mitmproxy

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt's example mitmproxy addons explicitly capture and print authorization tokens (even if truncated) and show header-modification examples that would cause secret values to be read and output verbatim, creating an exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The required runtime workflow involves intercepting and inspecting HTTP(S) traffic (requests/responses) from the target application via mitmproxy/mitmweb/mitmdump, and those response bodies and other free-form text are then available to the agent/LLM for analysis (outsider-authored content from external systems).

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

• Potentially malicious external URL detected (high risk: 0.90). The CI workflow in assets/ci-config-template.yml executes a remote install script at runtime using "curl -s https://raw.githubusercontent.com/aquasecurity/tfsec/master/scripts/install_linux.sh | bash", which fetches and directly executes remote code during skill runtime.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The prompt explicitly instructs actions that change system/network state and bypass security (modify /etc/hosts, require iptables/pf rules for transparent proxy, bind to privileged port 443, and use SSL unpinning tools), which require elevated privileges or circumvent protections.