add-signal

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The instructions require pasting sensitive values (captcha tokens, a phone number, and .env credentials) verbatim into CLI commands and files, which would force an LLM to handle and output secrets directly (exfiltration risk).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's setup fetches and installs a remote binary via curl from GitHub releases (calls to https://api.github.com/repos/AsamK/signal-cli/releases/latest and the subsequent https://github.com/AsamK/signal-cli/releases/download/.../signal-cli-...-Linux-native.tar.gz) and also pulls adapter code from the repository with git fetch/git show, which clearly fetch remote code that will be executed/installed and is required for the channel to run.