crypto-risk-monitoring
Pass
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill describes a legitimate cryptocurrency risk monitoring workflow. The tools listed are relevant to the stated purpose of tracking exposure and alerting users, and there are no signs of obfuscation, malicious commands, or credential harvesting.
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it ingests untrusted external data such as token names, prices, and account summaries.
- Ingestion points: External data is ingested via tool calls list_positions, get_price, and check_watches as described in SKILL.md.
- Boundary markers: No explicit instructions for delimiters or 'ignore embedded instructions' are present in the skill text.
- Capability inventory: The skill utilizes tools for messaging (send_message), generating artifacts (publish_artifact), and modifying risk parameters (adjust_risk_limits).
- Sanitization: No specific sanitization or validation logic is described for the content of the external data processed by the agent.
Audit Metadata