The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external sources, specifically phone call transcripts and SMS message content, which may contain adversarial instructions.
Ingestion points: Untrusted data enters the agent's context through webhook payloads (specifically the agent.message event described in references/detailed-guide.md) and API calls such as GET /v1/calls/{call_id} which returns the transcripts array.
Boundary markers: The skill instructions do not specify any delimiters (like xml tags or markdown blocks) or "ignore instructions within data" warnings to isolate these external transcripts from the agent's operating context.
Capability inventory: The skill allows the agent to perform network requests (curl to the AgentPhone API), manage account resources (buying or releasing phone numbers), and initiate communications (making calls or sending SMS) as documented in SKILL.md and references/detailed-guide.md.
Sanitization: No sanitization, validation, or filtering of the incoming transcript text or message body is required or described in the instructions.

agentphone