The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the npx agenticloops CLI utility to manage the lifecycle of recurring agents, including searching, installing, and running loops.
[REMOTE_CODE_EXECUTION]: The skill facilitates the installation and execution of agent configurations and prompts from external GitHub repositories. While it provides vetting guidelines such as checking for verifiable run receipts and source reputation, the capability to execute logic from arbitrary remote sources is a significant security consideration.
[PERSISTENCE]: The core functionality of the skill is to register 'recurring agents' that maintain a persistent presence in the environment by executing tasks on a set schedule (e.g., hourly, daily) or in response to specific events.
[INDIRECT_PROMPT_INJECTION]: The agents managed by this skill are intended to process external and potentially untrusted data from sources like the web or code repositories, creating a surface for indirect prompt injection.
Ingestion points: Data retrieved from external websites (research loops), GitHub repositories (triage loops), or other network resources during scheduled runs.
Boundary markers: The skill does not mandate the use of explicit delimiters or 'ignore instructions' markers for data processed within the agent prompts.
Capability inventory: Agentic loops can be configured with access to specific CLI tools, MCP servers, and network egress points, providing a range of capabilities that could be abused if an injection occurs.
Sanitization: The instructions for authoring loops do not include requirements or methods for sanitizing, validating, or escaping external content before it is interpolated into agent prompts.

loops