The Agent Skills Directory

[COMMAND_EXECUTION]: The skill mandates the use of shell commands to perform calculations. It specifically instructs the agent to use Bash to call python3 -c or bc to execute mathematical expressions. This pattern involves spawning subprocesses via the command line.
[DYNAMIC_EXECUTION]: The skill generates and executes code at runtime. It requires the agent to transform natural language requirements into Python code or shell scripts for execution to ensure calculation accuracy.
[PERSISTENCE]: The skill implements a manual persistence mechanism by writing research results to a local directory. It specifically targets the path memory/entities/[公司简称].md to store data across different sessions, allowing the agent to recall past research.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes data from untrusted external sources.
Ingestion points: Data is ingested via the web-search tool and the searchis-query tool (which reads internal documents and expert interviews).
Boundary markers: The instructions do not define specific delimiters or "ignore" instructions for the data retrieved from external searches.
Capability inventory: The skill possesses file writing capabilities (Write tool) and command execution capabilities (bash, python3).
Sanitization: There is no explicit instruction to sanitize or validate external content before it is used in calculations or written to the persistence memory.

company-research-analyst