market-regimes
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill relies on 'Recommended approach' directives that are automatically generated from external market data and observer agent logs, creating a surface for manipulating agent behavior.
- Ingestion points: Behavioral instructions in
SKILL.mdand pattern descriptions in.pattern_history.jsonare populated by an external 'Observer Agent' based on processed market samples. - Boundary markers: The skill content lacks delimiters or instructions for the agent to distinguish between its own internal logic and the potentially attacker-influenced data in the regime descriptions.
- Capability inventory: The strategy descriptions (e.g., 'SHORT-biased strategy', 'position sizing at 2% equity risk') imply the agent possesses capabilities to execute financial transactions and modify risk management parameters via external APIs.
- Sanitization: There is no evidence of filtering or validation for the generated strategy text to prevent the inclusion of malicious imperatives that could override intended trading logic.
Audit Metadata