The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands including git mv for moving files and rm for deleting obsolete or backup files. These actions are governed by the workflow's Phase 4 and 5, which mandate creating a proposal and obtaining explicit user approval before any changes are executed.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its ingestion of project file and directory names during the assessment phase.
Ingestion points: The skill reads the project's directory structure using tree and ls -la (documented in Phase 1 of SKILL.md).
Boundary markers: The workflow establishes a boundary by requiring a written "Proposal" in Phase 4 and manual "User Approval" in Phase 5 before execution, preventing automated obedience to injected instructions.
Capability inventory: The skill has the capability to move files, delete files, and commit changes to the repository via the Task tool as described in Phase 5 (SKILL.md).
Sanitization: The skill instructions do not specify any sanitization or validation logic for the file system metadata retrieved from the environment before it is processed by the agent.

declutter