The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches and executes the @call-e/cli NPM package via npx as a fallback execution route. This package belongs to the vendor's verified scope and is used to facilitate the core call-making functionality.\n- [COMMAND_EXECUTION]: Instructs the agent to create recurring jobs using system schedulers. The skill includes specific safeguards, such as requiring explicit user consent, providing cancellation instructions, and masking phone numbers to protect user privacy.\n- [PROMPT_INJECTION]: The skill processes user-supplied reminder messages that are stored and subsequently executed by an agent, presenting an indirect prompt injection surface.\n
Ingestion points: User-provided reminderMessage input processed by scripts/render-runtime-prompt.mjs and defined in SKILL.md.\n
Boundary markers: The references/runtime-prompt.md template utilizes fenced text segments and a checklist of mandatory runtime checks to guide agent behavior.\n
Capability inventory: The agent's actions are restricted to the verified one-off CALL-E call workflow, as detailed in references/calle-cli-bootstrap.md.\n
Sanitization: Structured data fields like cadence, time, and phone numbers are validated using regular expressions in scripts/validate-reminder-input.mjs.

call-reminder