The Agent Skills Directory

[PERSISTENCE]: The skill provides detailed instructions and configuration templates for creating launchd (macOS) and systemd (Linux) service files. These configurations are designed to ensure the GrepAI watch daemon starts automatically at system boot and remains active across sessions.
[PRIVILEGE_ESCALATION]: The troubleshooting section includes a command that uses sudo tee to modify the system's inotify watch limits on Linux. This action requires elevated administrative privileges to change kernel-level file system parameters.
[INDIRECT_PROMPT_INJECTION]: The skill indexes and monitors all source files in a project directory, which acts as a primary ingestion point for potentially untrusted data that could influence agent behavior.
Ingestion points: All source files matching supported extensions within the project directory.
Boundary markers: No explicit instructions to the LLM to ignore or delimit embedded instructions within the source code were identified.
Capability inventory: The skill manages background processes, creates system service files, and executes shell commands for daemon management.
Sanitization: No specific sanitization or validation steps for the content being indexed are described.
[COMMAND_EXECUTION]: The skill guides the agent to execute multiple shell commands for managing the daemon life cycle, including grepai, systemctl, launchctl, and tail.

grepai-watch-daemon