ae-user-metric-quantile
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where user-controlled inputs are directly interpolated into shell-based CLI commands.\n- Ingestion points: User-provided strings for event names, property names, and aggregation parameters are collected via the AskUserQuestion tool as described in the Prerequisite Information Collection section of SKILL.md.\n- Boundary markers: There are no instructions for the agent to use delimiters, quotes (other than the placeholders), or safety warnings to ignore instructions embedded within the user-provided names.\n- Capability inventory: The skill utilizes the ae-cli tool to perform sensitive operations including creating tags (analysis user-tag create), virtual properties (analysis-meta virtual-property create), and reports (analysis report create). These operations involve shell command execution and interaction with an external analytics server.\n- Sanitization: The command templates in references/commands.md show direct interpolation of placeholders (e.g., --tag-name <Step1_tag_name>) without explicit sanitization or escaping of shell meta-characters, which could be exploited for command injection if malicious names are supplied.
Audit Metadata