herdr-throwaway-repro

Pass

Audited by Gen Agent Trust Hub on Jul 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on shell commands to manage isolated Herdr sessions. It uses env to isolate environment variables and herdr CLI commands to create panes, run processes, and manage agents.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface because it is designed to ingest and act upon data generated within a test environment.
  • Ingestion points: The skill captures untrusted data via pane read, agent read, agent explain, and general API outputs (SKILL.md).
  • Boundary markers: There are no explicit instructions or delimiters provided to ensure the agent ignores or sanitizes instructions embedded within the captured terminal output.
  • Capability inventory: The agent is granted capabilities to execute commands (pane run), start new agents (agent start), and delete sessions (session delete) based on the data it reads.
  • Sanitization: The instructions do not specify any validation, escaping, or sanitization of the external content before it is processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 27, 2026, 10:22 AM
Security Audit — agent-trust-hub — herdr-throwaway-repro