The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts (lint_obsidian_mechanics.py and build_governance_indices.py) located in a sibling directory (../obsidian-notes-karpathy/scripts/) to automate linting and governance indexing tasks.
[PROMPT_INJECTION]: The skill processes content from various sources such as wiki/live/ and outputs/qa/ (ingestion points) which may contain untrusted instructions. It lacks explicit boundary markers or sanitization logic. The skill maintains capabilities for file modification and command execution (capability inventory), which could be targets for indirect injection attacks.

kb-health