droid-recon

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to run secret-hunting commands, "collect output" and embed grep/rg results and evidence into the incremental Markdown report (including a secrets table and "Evidence" snippets), which requires handling and potentially outputting secret values verbatim.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.70). Several of the URLs are not direct download links but indicate sensitive internal endpoints and embedded webhook/database endpoints (internal IP, dev/internal host, Firebase URL, Slack/Discord webhook patterns) which are high-risk indicators for information exposure or abuse and can be used to host or deliver malicious payloads.

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

• Third-party content exposure detected (medium risk: 0.30). Source files from the uploaded APK are attacker-controlled, and the required workflow explicitly decompiles/unzips and then runs ripgrep/grep over the resulting readable text from those artifacts (e.g., JADX output and apktool output, plus assets/config), which is free-form outsider-authored content.

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

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill explicitly instructs installing missing tools using "sudo apt install" (lines with sudo apt install), which requests elevated privileges and would modify the host system state.