openhands
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill makes extensive use of the
terminaltool to execute shell commands. This is the primary purpose of the skill to interface with the OpenHands CLI. - [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the
openhandstool usinguv tool install. The source repository is a well-known project (All-Hands-AI/OpenHands). - [INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it takes user-provided tasks (
-t 'TEXT') and project files from a workspace and feeds them into the OpenHands agent. The skill includes some capability to run the agent in--headlessmode which auto-approves actions, increasing the impact if an injection occurs within the project files. However, this is inherent to the function of a coding agent skill. - [CREDENTIALS_SAFE]: The skill provides instructions for managing API keys via environment variables (e.g.,
LLM_API_KEY=$OPENROUTER_API_KEY), which is a standard and safe practice for secret management in developer tools. It does not contain hardcoded secrets.
Audit Metadata