configuring-experiment-analytics

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill is purely instructional, providing guidance on using PostHog API tools and JSON schemas for experiment configuration. All referenced tools and documentation relate to the legitimate 'posthog' vendor environment and show no signs of malicious intent or unauthorized access patterns.
  • [INDIRECT_PROMPT_INJECTION]: The skill identifies a workflow where the agent ingests external project data to configure settings, creating a surface for potential indirect injection if project data is attacker-controlled.
  • Ingestion points: Event names and metric definitions retrieved via read-data-schema, experiment-get, and experiment-saved-metrics-list (SKILL.md).
  • Boundary markers: The instructions do not specify delimiters for data isolation, though they enforce strict JSON schema adherence for outputs.
  • Capability inventory: The agent can modify project state using experiment-update and experiment-saved-metrics-create based on the processed data.
  • Sanitization: The documentation emphasizes structural correctness and schema compliance rather than explicit string sanitization or filtering.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 01:06 AM
Security Audit — agent-trust-hub — configuring-experiment-analytics