open-dynamic-workflows
Warn
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructions direct the user to clone and install dependencies from a community repository (github.com/Suraj1235/open-dynamic-workflows) using
npm install. This involves downloading and running code from a source outside of established trusted organizations. - [REMOTE_CODE_EXECUTION]: The skill utilizes
npx odw-daemonandnpm run odw, which executes the code bundled within the community-provided package. This pattern facilitates the execution of remote code maintained by the repository author. - [COMMAND_EXECUTION]: A core feature of the skill is the execution of agent-generated code and shell commands. While this is the stated purpose of the workflow engine, it requires the agent to have high-privilege interaction with the local shell environment.
- [DATA_EXFILTRATION]: The skill manages sensitive model provider credentials (such as
ANTHROPIC_API_KEY) and performs network operations to reach these providers. While standard for AI agents, these actions constitute an attack surface for potential credential exposure or data exfiltration if the community engine is compromised.
Audit Metadata