speak-prod-checklist
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill incorporates an attack surface for indirect prompt injection by processing outputs from external tools and services, such as API health endpoints and status pages. This is a common characteristic of agent skills performing system verification.
- Ingestion points: Tool outputs from
curlcommands interacting with the Speak API, status endpoints, and Prometheus metrics inSKILL.mdandreferences/implementation-guide.md. - Boundary markers: The instructions do not include specific delimiters or instructions to ignore embedded commands within tool outputs.
- Capability inventory: The skill is scoped to use
Read,Write,Edit,Bash, andGreptools. - Sanitization: There is no explicit logic to sanitize or filter data retrieved from external API calls before it is presented to the agent context.
Audit Metadata