The Agent Skills Directory

[SAFE]: No malicious patterns or security risks detected. The skill's behavior is consistent with its stated purpose of document classification and compliance marking.
[COMMAND_EXECUTION]: The skill uses uv run and bash scripts to manage and execute its Python components. These are standard practices for environment management and do not involve unauthorized shell execution.
[EXTERNAL_DOWNLOADS]: The project dependencies (typer, httpx) are well-known, reputable libraries declared in pyproject.toml. There are no instances of remote script downloads or piped execution.
[DATA_EXFILTRATION]: The skill communicates with a local datalake-daemon service using a Unix Domain Socket (/run/user/1000/embry/datalake.sock) to fetch documents for internal scanning. There is no evidence of data being transmitted to external servers or untrusted domains.
[PROMPT_INJECTION]: The instructions do not contain any patterns suggesting attempts to override system prompts, bypass safety filters, or extract sensitive instructions.

cui-marker