crypto-daytrading
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data in the form of market bars (OHLCV), funding rates, and volatility metrics via the CCXT library or exchange APIs.
- Ingestion points: Market data pulled from external exchange APIs (mentioned in the daily operating loop in SKILL.md).
- Boundary markers: The skill does not specify the use of delimiters or 'ignore embedded instructions' warnings for the data ingested from APIs.
- Capability inventory: The skill has the capability to execute shell commands via the Coinbase CDP CLI and write to an immutable audit log.
- Sanitization: There is no mention of sanitizing or escaping the incoming market data before it is processed by the agent to determine trading regimes or signals.
- [COMMAND_EXECUTION]: The skill instructions include the use of the
CDP CLIfor order execution. While this is the intended primary purpose of the skill, it represents a capability that can be triggered by logic derived from external, untrusted inputs.
Audit Metadata