sm-batch-refresh

Pass

Audited by Gen Agent Trust Hub on Jun 25, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill integrates with external financial data tools, specifically the iFind MCP, to query real-time market pricing, financial statements, and shareholder data. These calls are essential for the skill's primary purpose.
  • [PROMPT_INJECTION]: Indirect prompt injection surface identified. 1. Ingestion points: The skill reads from local files (coverage_root/INDEX.md) and external financial data sources (news and announcements) provided by tools. 2. Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded instructions in the ingested news data. 3. Capability inventory: The skill possesses the ability to read and write to the local file system and execute data retrieval tools. 4. Sanitization: No sanitization or validation of the external content is performed before integration into reports.
  • [SAFE]: File system operations, including reading the coverage index and writing updated reports to archival directories, are localized and consistent with the stated functionality of maintaining a financial research coverage pool.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 25, 2026, 01:59 AM
Security Audit — agent-trust-hub — sm-batch-refresh