elite-developer

Pass

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill acknowledges the ingestion of untrusted external data, creating a potential attack surface.
  • Ingestion points: External data and web content accessed via BrowserOS (SKILL.md).
  • Boundary markers: The instructions explicitly state that external data must be treated as hostile.
  • Capability inventory: Includes terminal command execution via the rtk prefix and persistent file writing via the MemoryStore interface (SKILL.md).
  • Sanitization: Explicit instructions are provided to redact secrets and PII before sharing context with subagents.
  • [COMMAND_EXECUTION]: The skill utilizes a terminal execution proxy (rtk prefix) for tasks like running tests. This is a standard capability for a developer-oriented skill and is governed by strict operational modes described in the protocols.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 24, 2026, 01:35 PM
Security Audit — agent-trust-hub — elite-developer