The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The installation instructions require cloning a repository from an untrusted source: https://github.com/chiefautism/privacy-parser.
[REMOTE_CODE_EXECUTION]: The skill instructs the user to run uv pip install -e on local packages within the cloned, untrusted repository (./privacy-filter and ./pii_parser). This allows for arbitrary code execution during the installation phase.
[EXTERNAL_DOWNLOADS]: The skill documentation states that the first run automatically downloads a 3GB binary checkpoint (opf weights) to ~/.opf/privacy_filter/. The origin and integrity of this large binary are not disclosed, posing a high risk of malware delivery.
[DATA_EXFILTRATION]: The skill is designed to automatically detect and extract structured PII (Personally Identifiable Information). In the context of an AI agent, this provides a powerful tool for scanning large volumes of text to identify and isolate sensitive data spans.
[CREDENTIALS_UNSAFE]: The skill's label taxonomy specifically includes a secret category for identifying passwords, tokens, and API keys, and an account_number category for bank identifiers. This facilitates the automated discovery and potential theft of credentials from unstructured data.

privacy-parser-pii-extraction