The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a bundled script (content_humanize.py) using a vendor-specific CLI tool (claude-seo). This execution occurs when the user requests content humanization or watermark removal, allowing the agent to process local files and generate output drafts.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external sources, creating a potential vector for indirect instructions.
Ingestion points: Data is ingested from user-provided URLs and the content of draft files (e.g., draft.md) during the SEO audit and cleanup workflows.
Boundary markers: The instructions do not specify the use of delimiters or specific "ignore" instructions when the agent reads or processes the external content.
Capability inventory: The skill has the ability to execute shell commands, run Python scripts, and write to the local filesystem (-o cleaned.md).
Sanitization: The content_humanize.py script performs active sanitization by stripping zero-width codepoints, Unicode tag characters, and directional overrides. This reduces the risk of the agent being influenced by hidden or obfuscated instructions within the processed text.
[DYNAMIC_EXECUTION]: The skill employs dynamic execution by invoking a Python script at runtime to perform deterministic text normalization and phrasing replacements. This is used to implement the 'Draft Cleanup' functionality described in the instructions.

seo-content