The Agent Skills Directory

[SAFE]: The skill is implemented with a clear security-first approach.
Sanitization: The card_markup.py module implements a sophisticated whitelist-based sanitizer for Google Chat's specific markup requirements. It canonicalizes allowed tags (like bold, italics, and HTTPS links) and neutralizes all other markup by converting it into non-rendering HTML comments.
Secret Management: It avoids hardcoding credentials, instead reading the Google Chat webhook URL from a .env file. The documentation explicitly warns the user that the webhook URL is a secret and should be treated like a password.
Data Integrity: The skill uses Unicode lookalike characters for data-heavy fields (like project names and timestamps) to ensure they are rendered as text and cannot be misinterpreted as executable markup tokens.
Execution Safety: The bash wrapper provides a thin, secure interface that handles environment variables without using source, mitigating risks of arbitrary command execution from malformed configuration files.
Fail-Silent Behavior: The notification logic is designed to log warnings to stderr and exit gracefully on network failures, preventing external errors from blocking the agent's primary tasks.

notify-gchat-human-wait