The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the uv run command to execute a local tool (cleanuagent).
[EXTERNAL_DOWNLOADS]: The CLI tool employs a --sync-remote flag, which suggests it may download configuration or code updates from a remote server at runtime.
[PROMPT_INJECTION]: The skill includes instructions that attempt to constrain the agent's writing style. While functional, these instructions directly influence the core logic of the model's generation process.
[INDIRECT_PROMPT_INJECTION]: The skill is specifically designed to ingest and process untrusted data from web pages and external text sources.
Ingestion points: ContentCleaner.clean_url(...) and uv run cleanuagent url <URL> functions in SKILL.md indicate ingestion of arbitrary web content.
Boundary markers: There are no specific delimiters or 'ignore' instructions provided to the agent to prevent malicious content within the cleaned data from being interpreted as instructions.
Capability inventory: The skill has the ability to execute CLI commands and perform network requests to retrieve URL content.
Sanitization: While the tool aims to remove 'marketing jargon', it lacks security-focused sanitization to strip potential prompt injection attacks hidden in the scraped source material.

cleanuagent-sanitizer