google-analytics
Pass
Audited by Gen Agent Trust Hub on Aug 15, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill contains instructions that direct the agent to override its current context and adopt a workflow from a relative path (
../../analytics/google-analytics/SKILL.md). By commanding the agent to 'follow it as the active workflow', it attempts to dynamically change the operating instructions from an external source. - [INDIRECT_PROMPT_INJECTION]: The skill serves as a redirector, ingesting instructions from the local file system using path traversal patterns. This creates a vulnerability surface where the agent's behavior is governed by content located outside the immediate skill package.
- Ingestion points: Local file system access via relative path traversal (
../../). - Boundary markers: Absent; the agent is instructed to read the external file 'completely' without safety delimiters.
- Capability inventory: The skill itself contains no direct tool or command executions, but it delegates all capabilities to an external file.
- Sanitization: Absent; there is no validation of the content or the target file before the agent is told to follow it.
Audit Metadata