The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local shell commands to perform data verification and arithmetic tasks. It calls specific local scripts such as tools/financial_rigor.py and tools/report_audit.py, alongside the system date utility. These executions are constrained to the local repository environment and serve legitimate analytical purposes.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external financial data and user-provided parameters, creating an attack surface for indirect instructions.
Ingestion points: User request arguments ($ARGUMENTS), external financial filings, investor documents, and local portfolio files.
Boundary markers: The instructions do not explicitly define technical delimiters for user input interpolation, but they enforce logical boundaries by requiring all statements to be labeled as fact, estimate, or judgment.
Capability inventory: Shell execution for local Python tools, and file system access to read project data and write final reports to the reports/ directory.
Sanitization: The skill mandates a 'Research Discipline' that requires cross-checking data against at least two independent sources and passing an automated audit script (tools/report_audit.py) before a report is considered valid.

income-investment