airbyte-agent
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The
SKILL.mdfile contains a command to install the necessary CLI tool usingcurl -fsSL https://airbyte.ai/install.sh | bash. This URL belongs to the official domain of the skill's authoring organization and is a standard distribution method for the required software. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates reading data from third-party SaaS platforms (e.g., Slack, HubSpot, Salesforce) through the
connectors executecommand. This creates a surface for indirect prompt injection where instructions embedded in external data could influence the agent. - Ingestion points: Data is retrieved via the
connectors executeoperation described inreferences/connectors-execute.md. - Boundary markers: The skill does not provide specific instructions for using delimiters to isolate external data from the prompt.
- Capability inventory: The agent can list, read, create, update, and delete connector data using the
airbyte-agentCLI. - Sanitization: No explicit sanitization or filtering instructions are provided for the data fetched from external SaaS providers.
- [COMMAND_EXECUTION]: The skill's primary function is to drive the
airbyte-agentCLI. It provides comprehensive instructions for executing subcommands to manage workspaces, organizations, and connectors.
Audit Metadata