The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill demonstrates a vulnerability surface for indirect prompt injection due to its handling of untrusted data from the Inngest API.
Ingestion points: The skill retrieves and processes external data via endpoints like GET /runs/{runId}/trace (which includes function inputs and outputs) and GET /insights/events/schemas as documented in SKILL.md and references/rest-api-v2.md.
Boundary markers: The instructions lack specific guidance on using delimiters or "ignore previous instructions" markers when the agent processes these external data payloads.
Capability inventory: The skill has the capability to perform network operations using curl and execute local commands via npx inngest-cli.
Sanitization: While the instructions explicitly mandate the redaction of sensitive tokens and secrets, they do not provide measures for sanitizing incoming text payloads for potentially malicious instructions.

inngest-api