The Agent Skills Directory

[SAFE]: The skill operates entirely within the local development environment, primarily interacting with .claude/solutions/ and .claude/plans/ to manage institutional knowledge. All file operations (reading scratchpads, writing solution docs) follow standard developer workflows.
[COMMAND_EXECUTION]: The skill uses local shell commands like grep, mkdir, and ls to search and organize local documentation. These are used safely to extract context from recent work or to find existing solution docs to prevent duplicates.
[DATA_EXPOSURE]: The skill reads from local files such as .claude/plans/*/progress.md and git diff to gather context about recent bug fixes. This data is used solely to generate internal documentation and is not sent to external servers.
[EXTERNAL_DOWNLOADS]: The skill references local schema and template files stored within the plugin directory (${CLAUDE_PLUGIN_ROOT}). It does not perform network operations or download external code.
[PROMPT_INJECTION]: No evidence of prompt injection, role-play bypasses, or instructions to override safety filters was found. The instructions focus on structured data capture and validation against local schemas.

compound