monitor-experiments-consolidated
Pass
Audited by Gen Agent Trust Hub on Aug 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill uses established platform-specific tools under the Amplitude namespace to perform its stated purpose of experiment monitoring and reporting. The instructions focus on data retrieval, filtering, and summarization without any suspicious command execution, data exfiltration to external domains, or persistence mechanisms.- [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface that processes untrusted metadata from experiments and AI-generated feedback.
- Ingestion points: External data enters the context via the
Amplitude:search_amp_entities,Amplitude:use_amp_experiments, andAmplitude:use_amplitude_ai_feedbacktools defined inSKILL.md. - Boundary markers: There are no explicit delimiters or instructions provided to the model to ignore potential commands embedded within experiment titles or feedback descriptions.
- Capability inventory: The skill's capabilities are limited to read-only reporting; it does not utilize tools for file system writing, network requests to third-party domains, or shell execution.
- Sanitization: The skill relies on field-level extraction (fetching specific summary and validity flags) rather than string sanitization to mitigate risks from processed content.
Audit Metadata