The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting external data from web searches and user-provided arguments. Ingestion points: Untrusted data enters the agent context via the $ARGUMENTS variable and the results of the WebSearch tool during the global company scanning phase. Boundary markers: The skill instructions lack explicit delimiters or warnings to ignore embedded instructions found within external search results. Capability inventory: The skill possesses the Bash tool to execute local scripts and the Write tool to generate markdown reports in the home directory (~/[行业名]产业链投资研究报告.md). Sanitization: No sanitization or validation steps are defined for external content retrieved during the research process.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute local utility scripts and system commands to facilitate its research workflow. Evidence: The instructions require running the date command for temporal baselining and local Python scripts such as tools/financial_rigor.py and tools/report_audit.py for data processing and auditing.

industry-research