The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to handle Instantly event data, which can include user-generated content such as email bodies and lead replies. This introduces a theoretical surface for indirect prompt injection where malicious instructions within those emails could be interpreted by the agent.
Ingestion points: Webhook payloads containing lead and email content as specified in SKILL.md and references/official-docs.md.
Boundary markers: The instructions do not specify prompt-level delimiters, but they do mandate schema validation.
Capability inventory: The skill has access to WebFetch, Write, and Edit tools.
Sanitization: The skill explicitly requires schema validation, timestamp bounding, and redaction of sensitive fields.
[EXTERNAL_DOWNLOADS]: The skill utilizes official developer resources and packages from Instantly.
References: Targets developer.instantly.ai and api.instantly.ai for documentation and API access.
Packages: Mentions official tools @instantlyai/sdk and @instantlyai/cli for integration.

instantly-webhooks-events