webapp-testing
Warn
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script scripts/with_server.py executes arbitrary commands provided via the --server flag and the trailing command arguments. It specifically uses subprocess.Popen with shell=True to support complex shell operations like cd and &&, which increases the risk of command injection.- [DYNAMIC_EXECUTION]: The skill relies on running dynamically provided commands to start development servers and execution scripts. While intended for local testing, this pattern allows for the execution of any shell command provided as an argument.- [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection if the agent is instructed to test a malicious web application that contains instructions designed to manipulate the agent's actions or the commands it passes to the helper scripts.
- Ingestion points: scripts/with_server.py accepts command-line arguments --server and command which are executed.
- Boundary markers: None present to delimit or validate the commands being executed.
- Capability inventory: The skill can execute shell commands (subprocess.Popen, subprocess.run), write to the filesystem (/mnt/user-data/outputs/), and capture screenshots.
- Sanitization: No sanitization or validation is performed on the commands passed to scripts/with_server.py.
Audit Metadata