The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts to perform quality checks on the generated branding materials.
Evidence: python scripts/quality_gates/banned_word_check.py <file> and python scripts/quality_gates/four_us_score.py <file>.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources and local files to inform its branding analysis, creating a surface for indirect prompt injection.
Ingestion points: The skill reads content from MARKETING.md, CLAUDE.md, README.md, PROJECT.md, package.json, and content fetched from user-provided URLs.
Boundary markers: None identified. The instructions do not specify the use of delimiters or 'ignore embedded instructions' prompts when processing these inputs.
Capability inventory: The skill can execute local Python scripts via subprocess and write files (MARKETING.md) to the local project root.
Sanitization: None identified. The instructions do not mention filtering or escaping content retrieved from external URLs or project files before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill is designed to fetch and analyze external website content if provided with a URL by the user.
Evidence: 'If the user provides a URL, fetch and analyze the current site messaging.'

kai-brand