The Agent Skills Directory

[COMMAND_EXECUTION]: The scripts/content_inventory.py script executes the hugo binary using subprocess.run to retrieve site metadata. This is a controlled execution of a specific, well-known tool required for analyzing Hugo-based projects.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data via web research to map competitive landscapes and benchmarks.
Ingestion points: Web search results and fetched landing pages for competitors (defined in references/competitor-analysis.md and references/mckinsey-consultant.md).
Boundary markers: The skill enforces a strict 'Evidence contract' requiring inline tags like [Evidence: URL] or [Assumption] and a mandatory 'Sources and Assumptions' section.
Capability inventory: The skill performs local file reads, executes the hugo command via subprocess, and writes analysis reports to the local filesystem.
Sanitization: The skill relies on structured fact extraction and agent-level interpretation rather than dynamic execution of fetched content, reducing the risk of accidental instruction obedience.

write-strategy