learning-analyzer

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection during the analysis phase.
  • Ingestion points: Data retrieved from Supabase views (e.g., v_common_patterns, v_feedback_summary) is interpolated directly into the AI analysis prompt via the {query_results} placeholder.
  • Boundary markers: The prompt template does not utilize clear delimiters or instructions to distinguish between the analysis instructions and the untrusted data being analyzed.
  • Capability inventory: While the analysis skill primarily outputs a report, it integrates with a 'Self-Improve' workflow (/learn:apply --auto) that suggests the agent may have the capability to modify configurations or rules based on the analysis output.
  • Sanitization: There is no evidence of sanitization or validation of the data retrieved from the database before it is processed by the LLM.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 05:11 PM