The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests and processes untrusted data from inbound messages and webhooks, creating a potential surface for indirect prompt injection attacks.
Ingestion points: The inboundMessage webhook event contains a text field and media metadata originating from external senders, and the List opt-outs operation retrieves user-supplied from numbers and keyword strings.
Boundary markers: The skill instructions do not specify the use of delimiters or clear separation between external message content and agent instructions.
Capability inventory: The skill allows the agent to send SMS/MMS, schedule messages, and perform administrative actions like RegenerateSecret for messaging profiles.
Sanitization: The implementation examples do not include logic for sanitizing or validating inbound message content before it is processed by the agent.
[SAFE]: The skill utilizes os.Getenv("TELNYX_API_KEY") in its setup example, adhering to the best practice of avoiding hardcoded credentials.
[SAFE]: The skill explicitly instructs users to verify webhook signatures using Ed25519, which helps protect against request spoofing and ensures data integrity.
[SAFE]: Dependencies are fetched from the official vendor repository github.com/team-telnyx/telnyx-go, which is a recognized and expected source for this integration.

telnyx-messaging-go