The Agent Skills Directory

[SAFE]: The analyzed skill consists of informational reference guides and boilerplate Python scripts. No evidence of malicious intent, data exfiltration, or persistence mechanisms was found.
[COMMAND_EXECUTION]: The skill instructions provide commands to run local Python scripts (bundle_analyzer.py, component_generator.py, frontend_scaffolder.py). These scripts were audited and found to contain only skeleton logic using standard libraries (argparse, pathlib, json). They do not execute arbitrary shell commands or perform dangerous system operations.
[EXTERNAL_DOWNLOADS]: The documentation mentions standard dependency installation via npm install and pip install, which is expected for the described development workflow. No suspicious or unverified third-party packages or remote scripts are explicitly referenced or downloaded.
[DATA_EXPOSURE]: No hardcoded credentials, API keys, or attempts to access sensitive system files (e.g., .ssh, .aws) were detected in any of the skill files.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze and process user-provided source code projects.
Ingestion points: Project paths provided as arguments to the Python scripts.
Boundary markers: Absent in the current skeleton implementations.
Capability inventory: Scripts currently only validate paths and print status; no file-writing or network capabilities are implemented.
Sanitization: Not applicable as the scripts do not yet process file content.
Assessment: While this represents a potential attack surface if the scripts are extended to read and output file content to the agent, the current implementation is safe.

senior-frontend