testing-in-production
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions suggest reading from a project-specific context file to streamline discovery questions, which introduces a potential vector for indirect prompt injection if that file contains untrusted content.
- Ingestion points: The
SKILL.mdfile instructs the agent to check.agents/qa-project-context.mdfor pre-answered project details. - Boundary markers: There are no specific instructions or delimiters provided to the agent to treat the content of the project context file as untrusted data or to ignore embedded instructions within it.
- Capability inventory: The associated code patterns in
references/patterns.mddemonstrate capabilities for performing network requests and interacting with external APIs (using@playwright/testand feature flag provider clients). - Sanitization: The instructions do not define any validation, filtering, or sanitization steps for the data ingested from the project context file.
Audit Metadata