The Agent Skills Directory

[SAFE]: The skill consists exclusively of markdown documentation and natural language instructions for analyzing web forms. It does not include any executable code, shell scripts, or automation logic.
[NO_CODE]: No programming languages (Python, JavaScript, etc.) or CLI commands are utilized within the skill definition.
[CREDENTIALS_SAFE]: The content was scanned for hardcoded secrets, tokens, and API keys; none were found. Standard advice regarding privacy statements and security badges is provided as a best practice for form design, not as an exfiltration vector.
[DATA_EXPOSURE_SAFE]: The skill does not request access to sensitive local directories (like .ssh or .aws) and does not perform any network operations to send data externally.
[REMOTE_CODE_EXECUTION_SAFE]: There are no patterns involving the download or execution of remote scripts (e.g., curl | bash) or the installation of third-party packages.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process user-provided information about forms, it lacks the technical capabilities (file writes, network calls, subprocess execution) required to turn an injection into a security breach. The threat surface is minimal and results in a safe assessment for this category.

form-cro