The Agent Skills Directory

Prompt Injection (LOW): Indirect Prompt Injection Surface. The skill ingests untrusted data from the financial_search tool (SKILL.md, Step 1) which could contain malicious instructions. Mandatory Evidence: (1) Ingestion points: financial_search tool output. (2) Boundary markers: Absent. (3) Capability inventory: Limited to data extraction and math; no file-writing, network-sending, or command-execution capabilities detected. (4) Sanitization: Absent.\n- Data Exposure & Exfiltration (SAFE): No hardcoded credentials, API keys, or access to sensitive file paths were detected.\n- No Code (SAFE): The skill consists entirely of markdown documentation and reference data files. No executable scripts or binary files are included.

dcf-valuation