The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data in the form of source code from the project environment.
Ingestion points: Source files are read during the analysis phase as described in the 'Read and Analyze' section of SKILL.md.
Boundary markers: There are no defined delimiters or instructions to the agent to ignore instructions embedded within the source code comments or strings.
Capability inventory: The skill possesses the capability to read any file in the repository, write markdown files to the docs/explain/ directory, and instantiate parallel Task agents for processing.
Sanitization: The skill does not implement any sanitization or filtering of the source code content before analysis.
[COMMAND_EXECUTION]: The skill utilizes shell-level commands to determine its operational scope.
Evidence: The workflow uses git diff --cached --name-only to identify staged files and Glob operations to perform directory discovery when the --all flag is used.

explain