The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local bash script scripts/manage-primary-channel.sh to perform atomic updates to the primary-channel.json configuration file. The script validates input and uses jq with proper argument passing to ensure data integrity.
[COMMAND_EXECUTION]: Directly uses jq in the instructions to append entries to ~/trusted-contacts.json. The commands use the --arg flag, which is a security best practice to prevent JSON injection by treating user-provided values as strings.
[SAFE]: Implements a restrictive identity recognition model. It instructs the agent to 'silently ignore' messages from unknown senders in group rooms and specifically forbids sharing API keys, tokens, or management capabilities with 'Trusted Contacts'.
[SAFE]: No remote dependencies or external code execution patterns were detected. All scripts and references are local to the skill's directory structure or known internal paths.

channel-management