The Agent Skills Directory

Prompt Injection (SAFE): The documentation and quick reference guide contain standard instructional language and legal framework summaries. No patterns of instruction override, safety filter bypass, or system prompt extraction were detected.\n- Data Exposure & Exfiltration (SAFE): No hardcoded credentials, sensitive file paths (such as SSH keys or environment files), or unauthorized network communication patterns were found in the provided markdown or JSON data files.\n- Indirect Prompt Injection (SAFE): The skill defines a surface for processing JSON data (e.g., sample_commercial_expropriation.json). Analysis of the sample data confirms it consists entirely of expected numeric and descriptive fields related to property valuation, with no embedded instructions or malicious payloads. Ingestion point: JSON data files; Boundary markers: None (standard JSON); Capability inventory: Referenced local script execution; Sanitization: None documented.\n- No Code Detected (NO_CODE): The core logic file (expropriation_calculator.py) and the skill definition (SKILL.md) are referenced in the documentation but not included in the provided set of files, precluding the detection of runtime or execution-based vulnerabilities in the current scope.

expropriation-compensation-entitlement-analysis