The Agent Skills Directory

[Secure Archive Extraction]: The extract_solution.py script incorporates several security checks to handle untrusted ZIP files safely. It enforces limits on the number of members, uncompressed size, and individual file sizes to prevent denial-of-service (zip-bomb) attacks. Additionally, it rejects path traversal attempts and absolute paths to ensure files are only written to the intended staging directory.
[Credential Safety]: The analyze_solution.py script scans for sensitive keywords like 'password' or 'token' to identify potential security risks in the assessed automation. The implementation is designed to flag the presence and category of these keywords without capturing or exposing the actual secret values, maintaining the confidentiality of credentials within the source artifacts.
[Output Sanitization]: The reporting logic in render_report.py sanitizes data extracted from the solution artifacts before including it in the final Markdown report. By escaping HTML and Markdown characters, it prevents formatting-based injection and ensures the report is rendered safely by the agent or host platform.
[Minimal Footprint]: The included scripts rely exclusively on the Python standard library, avoiding risks associated with third-party dependencies or external package registries. The processing is performed in a localized, temporary directory to minimize the impact on the host system.

power-automate-desktop-assessment