herdr
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests terminal output which may contain adversarial instructions.
- Ingestion points: Terminal data is read via
herdr pane readandherdr agent readinSKILL.md. - Boundary markers: Absent; there are no instructions to use delimiters for the read output.
- Capability inventory: The skill can execute shell commands (
herdr pane run), orchestrate agents (herdr agent start), and manage remote sessions via SSH (herdr --machine). - Sanitization: Absent; output is read directly into the agent's processing flow.
- [COMMAND_EXECUTION]: The skill's primary function is to execute shell commands via the
herdrCLI, allowing for the creation and control of terminal processes and sub-agents.
Audit Metadata