The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected. The skill is designed with a 'security-by-design' approach, incorporating mandatory confirmation gates for all operational actions.
[COMMAND_EXECUTION]: The skill references the use of git worktree for repository management and rg (ripgrep) for context discovery. These commands are standard for the tool's purpose and are only proposed as part of a plan that requires user confirmation before execution.
[PROMPT_INJECTION]: The skill is subject to potential indirect prompt injection as it reads repository documentation and source code to inform its planning. This is mitigated here by a mandatory human-in-the-loop review process. (1) Ingestion points: System context discovery via docs/INDEX.md and repository-wide searches using rg. (2) Boundary markers: Employs a specific output contract and structured plan template to maintain response integrity. (3) Capability inventory: Proposes git commands and file system changes which are documented in SKILL.md. (4) Sanitization: Requires explicit user approval ('Confirm this plan for execution?') before any proposed tasks are implemented.

feature-implementation-planner