ae-first-event-builder
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a command-line interface tool (
ae-cli) to manage data assets and execute SQL queries. This includes creating and deleting user tags and virtual events, which are standard operations for the tool's intended purpose. Execution is governed by strict workflows that require user verification of project settings and permissions. - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface where user-provided strings (event names and property labels) are interpolated into SQL templates and shell commands.
- Ingestion points: User-provided behavior names and properties are gathered in the 'Identify Intent' phase (Step 1).
- Boundary markers: The workflow enforces a 'Pre-operation confirmation' rule (Step 3), which requires the agent to display the full execution plan to the user for approval before running any commands.
- Capability inventory: The skill utilizes
ae-clifor database operations, asset management, and ad-hoc analysis as documented inoperations.md. - Sanitization: The skill relies on the 'plan-level' confirmation step to allow human review of the generated commands before execution, mitigating the risk of malicious injection through parameter values.
Audit Metadata