analyze-experiment-consolidated

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill acts as a legitimate extension for Amplitude, using dedicated MCP tools (Amplitude:use_amp_experiments, Amplitude:query_amplitude_data, etc.) to perform data-driven analysis. All actions are scoped to the intended functionality of experiment interpretation.- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data retrieved from external Amplitude entities such as experiment descriptions, metric names, and user feedback themes.
  • Ingestion points: Untrusted data enters the context via Amplitude:get_from_url, Amplitude:use_amp_experiments, and Amplitude:use_amplitude_ai_feedback as described in Steps 0, 1, and 7 of SKILL.md.
  • Boundary markers: Absent; there are no explicit instructions or delimiters used to warn the agent to ignore instructions that might be embedded in the retrieved experiment metadata.
  • Capability inventory: The skill's capabilities are limited to data retrieval, taxonomy searching, and structured reporting. It does not possess high-risk capabilities such as arbitrary command execution, file system modification, or external network exfiltration to non-vendor domains.
  • Sanitization: The instructions do not specify any sanitization or filtering for the retrieved text strings before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 05:11 PM
Security Audit — agent-trust-hub — analyze-experiment-consolidated