The Agent Skills Directory

[PROMPT_INJECTION] (SAFE): No direct prompt injection or instructions to bypass safety protocols were identified in the skill instructions or code snippets.
[EXTERNAL_DOWNLOADS] (SAFE): The skill utilizes the standard Python packages dspy and pydantic from public registries. No unauthorized remote code execution or suspicious script downloads were found.
[DATA_EXFILTRATION] (SAFE): No hardcoded credentials, API keys, or access to sensitive file paths (such as SSH keys or environment files) were detected.
[COMMAND_EXECUTION] (SAFE): The code snippets are restricted to data processing and modeling; no functions for system command execution, such as os.system or subprocess.run, are used.
[INDIRECT_PROMPT_INJECTION] (SAFE): Although the skill is designed to parse untrusted text data (emails, invoices, resumes), it does not expose any high-risk capabilities like file system modifications or network exfiltration, rendering the risk of indirect injection negligible in this context.

ai-parsing-data