The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/invoke.py uses subprocess.run to execute the kntnt.py engine via uv run. This is used to pass the user's input to the management component that performs the requested explanation.
[DYNAMIC_EXECUTION]: The skill implements a mechanism in scripts/invoke.py to dynamically locate its own management engine across a wide variety of potential installation paths (e.g., in subdirectories of ~/.claude/skills, ~/.cursor/skills, ~/.codeium/windsurf/skills, etc.). While this involves constructing and executing paths at runtime, it is clearly intended for cross-platform compatibility within the vendor's ecosystem.
[INDIRECT_PROMPT_INJECTION]: The skill ingests raw user input from the command line and passes it directly to the underlying engine through standard input, creating a surface for indirect prompt injection.
Ingestion points: User input provided as arguments to the /explain command is captured and passed to the subprocess in scripts/invoke.py.
Boundary markers: The skill documentation describes a -- separator to distinguish between formal instructions and contextual guidance, which can help structure inputs.
Capability inventory: The skill has the capability to execute local scripts and interact with the filesystem via the uv tool and Python's subprocess module.
Sanitization: No explicit input sanitization or validation logic is visible in the provided script files before the data is passed to the execution engine.

explain