The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes shell commands using the npx agenticloops and 5dive loop CLI tools to search for, validate, and manage recurring agents.
[PERSISTENCE]: The skill's primary function is to register recurring, scheduled tasks (loops) on the host system using cron-style triggers (e.g., 5dive loop install ... --cron="..."). This establishes persistence for automated agent runs.
[REMOTE_CODE_EXECUTION]: The skill facilitates the installation and execution of external 'loops' from a remote directory. These loops contain executable skills and prompts that are run by the agent harness.
[EXTERNAL_DOWNLOADS]: Fetches agent metadata and installation files from vendor-associated sources, specifically agenticloops.dev and the 5dive-ai/loops GitHub repository.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect injection by processing data from an external directory of agents.
Ingestion points: Data ingested from npx agenticloops find results, which include author-provided descriptions and loop names.
Boundary markers: The instructions do not specify formal delimiters for external data but provide explicit vetting steps for the agent to perform.
Capability inventory: The agent has capabilities for command execution, task scheduling, and remote package installation.
Sanitization: The skill mitigates risks by instructing the agent to vet the requires block (identifying requested CLI tools, secrets, and network egress) and to check source reputation before recommending installation.

find-loops