The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes Python scripts to perform Trello API operations. It utilizes Python heredocs with quoted delimiters ('PYEOF') to ensure that the script content is not subject to shell expansion or interpolation, providing a secure method for executing complex logic.
[EXTERNAL_DOWNLOADS]: Communicates with api.trello.com. This is a well-known service required for the skill's primary task of managing Trello boards and does not involve downloading untrusted executable code.
[SAFE]: Includes explicit instructions to avoid shell-based tools like curl when handling user-provided text. It recommends using Python's urllib.parse.urlencode to properly sanitize data, effectively mitigating risks of command injection or data corruption from special characters.
[SAFE]: Evaluated for indirect prompt injection surface (Category 8):
Ingestion points: User-provided inputs for card titles, descriptions, and checklist items.
Boundary markers: None explicitly defined in the instruction templates for user data interpolation.
Capability inventory: Network requests via Python's urllib, local file reads for board references, and subprocess execution for Python scripts.
Sanitization: The skill architecture favors Python string handling and URL encoding over shell command interpolation, which significantly reduces the risk of malicious user input influencing the execution flow.

hadl-board