The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill provides a run_python tool designed to execute arbitrary Python code within the Abaqus kernel via a TCP connection to 127.0.0.1:48152. This allows for runtime script assembly and execution, which can be misused if the agent is directed to execute unintended commands.
[COMMAND_EXECUTION]: Several tools, including submit_job and set_workdir, interact with the host system's file system and process management capabilities to manage simulation workflows and environment configuration.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes data from external files that may be influenced by third parties.
Ingestion points: The tools inspect_odb and monitor_job_status read content from external ODB files and job diagnostic logs (.sta and .msg files).
Boundary markers: No explicit delimiters or instructions are provided in SKILL.md to prevent the agent from following instructions potentially embedded within the simulation results or logs.
Capability inventory: The agent has access to sensitive capabilities including run_python (code execution), submit_job (process spawning), and set_workdir (file system manipulation).
Sanitization: The instructions do not specify any validation or sanitization routines for data extracted from result files before it is processed by the agent or used to influence subsequent tool calls.

socket-bridge