alphagbm-options-strategy
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external market data from API responses and local mock files, which presents a surface for indirect prompt injection. \n
- Ingestion points: Data is received from the AlphaGBM API (at
https://alphagbm.zeabur.app) and read from local files matchingmock-data/*.jsonas defined inSKILL.md. \n - Boundary markers: The instructions in
SKILL.mddo not specify delimiters or explicit instructions to isolate the ingested JSON data from the agent's core logic. \n - Capability inventory: The skill performs network communication (HTTP GET and POST requests) and file system access (reading mock data) as described in
SKILL.md. \n - Sanitization: There is no mention of sanitization, schema validation, or data filtering for the incoming external JSON content. \n- [EXTERNAL_DOWNLOADS]: The skill communicates with the AlphaGBM API at
https://alphagbm.zeabur.appto retrieve option strategy templates and market data. This endpoint is an official resource for the skill's primary functionality and is associated with the skill author's vendor infrastructure.
Audit Metadata