The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/with_server.py executes arbitrary shell commands provided via the --server and test command arguments using subprocess.Popen and subprocess.run. This provides a direct interface for executing shell-level operations.\n- [DYNAMIC_EXECUTION]: The tool employs subprocess.Popen(cmd_str, shell=True) in scripts/with_server.py (line 159), which allows for the dynamic execution of strings that can include shell metacharacters. This significantly increases the risk of command injection if the input strings are not strictly validated.\n- [REMOTE_CODE_EXECUTION]: The tool is a high-privilege execution vector. As identified in the skill's own security documentation (ASI05 RCE Guard) in SKILL.md, assembling command strings from untrusted external sources can lead to arbitrary code execution.\n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection attacks because it lacks technical sanitization for ingested data. \n
Ingestion points: CLI arguments in scripts/with_server.py (--server, command).\n
Boundary markers: None present in the script or instructions to distinguish between trusted instructions and potentially malicious data within the command strings.\n
Capability inventory: Full shell access through subprocess execution in scripts/with_server.py.\n
Sanitization: No validation, escaping, or filtering is performed on the command strings before execution.

webapp-server-runner