product-analytics
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill utilizes the official Microsoft Clarity MCP server (@microsoft/clarity-mcp-server) via npx and a hosted PostHog MCP server (https://mcp.posthog.com/mcp) to facilitate data reading.- [COMMAND_EXECUTION]: Provides detailed instructions for users to configure the agent's environment, including the addition of MCP servers and the setup of environment variables.- [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: Untrusted analytics data (such as page paths, metrics, and session information) is brought into the agent's context through Clarity and PostHog MCP servers as described in SKILL.md.
- Boundary markers: While the skill instructs the agent to replace specific IDs in addresses with placeholders (e.g., :id), it does not specify the use of formal delimiters or explicit instructions to ignore potentially embedded malicious content in other data fields.
- Capability inventory: The agent is empowered to weigh ingested data against hypotheses and write findings to repository notes.
- Sanitization: The skill employs identifier masking for privacy; however, it lacks comprehensive sanitization or filtering for arbitrary text fields (like page titles or custom event properties) that could contain malicious prompts.
Audit Metadata