The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill explicitly advises storing user credentials, including passwords and role levels, in plaintext within a local directory named pentest-data/. This represents a high risk of credential exposure if the local environment is compromised.
[EXTERNAL_DOWNLOADS]: The installation process involves cloning code from a non-trusted GitHub repository (github.com/HeaSec/Pentest-Lyan.git) and installing third-party dependencies using pip. The source is not verified as a trusted vendor.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves downloading and analyzing external, untrusted content (JavaScript logic, API endpoints) from target websites. This data is fed back into the agent's context for analysis and decision-making without documented sanitization or boundary markers, creating a high vulnerability to malicious instructions hidden in target web applications.
[COMMAND_EXECUTION]: The skill frequently executes shell commands for environment setup (npm install, pip install, git clone) and runs dynamic utility scripts like render_docx.py at runtime.

pentest-lyan-web-security-testing