gsd-validation
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes potentially untrusted data from browser testing and system logs, creating an attack surface for indirect prompt injection.\n
- Ingestion points: The workflow ingests data from
gsd-logs-collectandgsd-playwright-execute(which collects traces, screenshots, and logs) as defined in SKILL.md.\n - Boundary markers: A mandatory Human-In-The-Loop (HITL) gate is described in
references/gate-policy.md, requiring user intervention if flow gaps are detected before runtime tests proceed.\n - Capability inventory: The skill interacts with runtime tools for process lifecycle management (
gsd-server-start,gsd-server-stop-owned) and browser-based testing (gsd-playwright-execute).\n - Sanitization: The skill instructions do not specify any sanitization, filtering, or validation schemas for the data collected from logs or browser environments.
Audit Metadata