octav-api

Warn

Audited by Gen Agent Trust Hub on Feb 15, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The installation instruction npx skills add Octav-Labs/octav-api-skill refers to a third-party GitHub repository. Since Octav-Labs is not on the trusted organization list, the package content is unverifiable and requires a manual audit to ensure no malicious code is included.
  • PROMPT_INJECTION (LOW): The skill processes external data from blockchain networks, making it vulnerable to Indirect Prompt Injection. Ingestion points: Transaction history, DeFi protocol positions, and token metadata fetched via the Octav API. Boundary markers: Absent from documentation; no evidence of delimiters used to isolate untrusted data from the agent prompt. Capability inventory: Primarily data querying, tracking, and analysis (read-only). No transaction signing or system-level write capabilities are described. Sanitization: Documentation does not specify any sanitization or filtering of external transaction metadata.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 07:58 AM