The Agent Skills Directory

[SAFE]: No malicious patterns detected.
[DATA_EXPOSURE]: The skill includes a robust SENSITIVE_PATH filter in scripts/hook-lib.mjs designed to prevent the agent from accidentally reading or processing sensitive files such as .env, private keys (id_rsa, .pem), and credential files.
[COMMAND_EXECUTION]: The skill executes local commands such as node for its own helper scripts and optionally spawns established CLI tools like claude or codex to assist with automated code edits. These operations are scoped to the user's project and intended for functional tasks.
[EXTERNAL_DOWNLOADS]: The skill performs a lightweight version check against https://impeccable.style to notify the user of updates. This is a standard practice for maintainability and does not involve executing remote code.
[INDIRECT_PROMPT_INJECTION]: The skill has a large attack surface as it processes user-provided source code and design documentation. However, it includes explicit instructions for the agent to treat data as literal content rather than instructions, and its automated scanning logic is deterministic.

impeccable