The Agent Skills Directory

[DATA_EXFILTRATION]: The skill accesses '~/Library/Messages/chat.db', which contains the user's entire private iMessage history, presenting a significant privacy and data exposure risk.
[CREDENTIALS_UNSAFE]: Instructions for creating a local signing certificate use a hardcoded password ('imsg123') within the 'openssl' and 'security import' commands, which is a poor security practice for identity management.
[COMMAND_EXECUTION]: The skill performs sensitive system-level modifications, including using 'security add-trusted-cert' to add a self-signed certificate to the system's trusted root store, potentially compromising the integrity of the machine's certificate validation.
[COMMAND_EXECUTION]: The skill directs the user to grant Full Disk Access (FDA) to a custom binary via System Settings, which bypasses macOS's Transparency, Consent, and Control (TCC) security framework for that application.
[COMMAND_EXECUTION]: The skill relies on the execution of local shell scripts ('build-local.sh' and 'install-rpc-app.sh') whose contents are not defined within the skill and could execute unverified or dangerous code.
[COMMAND_EXECUTION]: The skill implements a persistence mechanism by installing a launchd agent ('~/Library/LaunchAgents/com.joel.imsg-rpc.plist') to ensure the daemon runs automatically across user sessions.

imsg-rpc