new-hero-insight
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates as a specialized data analysis tool with a strict, rule-based interaction flow. It enforces pre-execution rules that require metadata retrieval and field matching before any analysis is performed.
- [DATA_EXPOSURE]: The skill is designed to interact with internal ThinkingData database tables (
ta.v_event_projectIDandta.v_user_projectID). This access is strictly scoped to the primary purpose of game data analysis and does not involve exfiltration to external or unknown domains. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from game event logs (e.g., JSON fields containing hero lineups).
- Ingestion points: Data is ingested from the ThinkingData event and user tables via SQL queries.
- Boundary markers: The skill uses a strict validation phase ("Field Matching Confirmation Checklist") to ensure data structure, though it lacks explicit delimiters for raw field content during interpolation.
- Capability inventory: The agent possesses SQL generation and JSON parsing capabilities.
- Sanitization: The skill relies on fixed statistical calibrations and quantified thresholds to mitigate the impact of malformed or malicious data inputs.
- [COMMAND_EXECUTION]: The skill generates SQL queries for database interaction. These operations are governed by fixed templates and statistical rules provided in the instructions, minimizing the risk of arbitrary command injection.
Audit Metadata