The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses local system tools such as xmllint and rg (ripgrep) to perform security and compatibility checks on the generated HTML files before they are presented to the user.
[PROMPT_INJECTION]: The skill processes untrusted user-provided Markdown input to create HTML files that are automatically opened in a browser, creating an attack surface for indirect prompt injection.
Ingestion points: User-provided Markdown files and raw text inputs as described in SKILL.md.
Boundary markers: The skill relies on specific role-scoping instructions (limiting the agent to formatting only) rather than explicit delimiters for the processed content.
Capability inventory: The skill can write files to the local filesystem, invoke a browser to preview generated HTML, and execute shell commands for output validation.
Sanitization: The skill implements mandatory HTML character escaping and utilizes rg to detect and block forbidden elements including <script> tags, external http links, and CSS @import rules.

dbs-wechat-html