The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a local Python script (scripts/conversion_guard.py) to snapshot and verify the structure of Markdown files. This script uses standard libraries to extract headings, links, and list counts to ensure that the AI's editing process does not accidentally remove or weaken the article's commercial intent. The execution is limited to local file system operations within the user's project directory.
[INDIRECT_PROMPT_INJECTION]: As the skill is designed to process and refine user-provided marketing drafts, it possesses an attack surface for indirect prompt injection. This risk is inherent to any tool that processes external text data. The skill mitigates this by providing the agent with a rigid workflow, a 'non-negotiable rule' for preserving specific content types, and a requirement for explicit user approval before making strategic changes.
Ingestion points: Markdown draft files are read and processed by the agent and the Python helper script (Workflow Step 3 and 5).
Boundary markers: The skill uses explicit HTML comments () to define sections of text that the agent and verification script must treat as immutable.
Capability inventory: The skill has the ability to read and write local files and execute a bundled Python script for structural verification.
Sanitization: No specific input sanitization is implemented for the content of the Markdown drafts, as the skill's purpose is to edit the provided text.

remove-ai-slop-without-removing-ctas-and-good-marketing-copy