prophet-adversarial-auditor
Pass
Audited by Gen Agent Trust Hub on Mar 21, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect Prompt Injection surface. * Ingestion points: Historical market data and submission events enter the context via
load_run_historyas defined in the skill workflow. * Boundary markers: The skill does not explicitly define boundary markers or provide instructions to ignore embedded commands when processing ingested market data. * Capability inventory: The agent has network access toapp.prophetmarket.aiandapi.serendb.com(vendor resources) viaurllib.request, and database write access viapsycopg. * Sanitization: No explicit sanitization or filtering of the ingested external data is performed in the main runtime script before processing. - [DATA_EXFILTRATION]: The skill performs network requests to
app.prophetmarket.ai(the target auditing service) andapi.serendb.com(the storage infrastructure associated with the vendor). These operations are required for the skill's primary function and do not target unauthorized or suspicious domains. - [COMMAND_EXECUTION]: The script executes SQL DDL statements from a local file (
serendb_schema.sql) to initialize the storage schema. This is standard setup behavior for establishing the necessary database tables in the user's provisioned storage instance.
Audit Metadata