Security audit
Security checks for vulnerabilities and agentic risk
This skill is a local foreclosure valuation workbook generator with disclosed finance risks and no evidence of hidden data access, persistence, or remote execution.
Install only if you want a China-focused foreclosure/distressed-asset valuation helper. Treat its output as a spreadsheet model that depends on your own assumptions and due diligence, not investment advice; verify market comparables, tax exposure, occupancy/eviction risk, and any JSON input source before relying on the XLSX.
The trigger keywords are very broad and include common finance terms like '风险', '多少钱', '能买吗', and '评估', which can cause accidental invocation in unrelated conversations. Overbroad activation increases the chance the skill runs on unintended user input, potentially causing inappropriate financial guidance, unexpected file generation, or unwanted processing of sensitive deal information.
The activation conditions are phrased at a high level and cover broad investment scenarios such as '要不要买、出多少' and '客户问不良资产投资逻辑', without clear exclusion criteria. In an agent setting, ambiguous routing rules can misfire, causing the skill to engage outside its intended scope and produce high-stakes valuation output where a more appropriate workflow, safety check, or specialized tool should be used.
Lines L072-L073 and the manifest description say this skill outputs a 5-sheet XLSX and includes bid-ceiling reverse solving. But L084 says bid-cap reverse solving is part of a separate full version, creating a direct mismatch in what the skill claims to provide versus what this documented version actually includes.
This JSON file contains only a high-level Chinese description of the template's purpose and usage, but no explicit activation conditions, trigger phrases, or exclusion conditions. For manifest-scope files, that can make it unclear when the associated skill should or should not be invoked.
This JSON sample uses Chinese-only values for deal names, asset types, and cities throughout, indicating a fixed language/locale assumption. Because SQP-3 applies to all file types, a hardcoded locale without user choice or an explicit region-specific justification can violate language/locale policy expectations.
No suspicious patterns detected.
