OxideTrack

Blueprint execution skill. Implements phases from BLUEPRINT.md using a strict TDD loop. Use when a BLUEPRINT.md exists and you are implementing tasks from the roadmap. Invoke for phase execution, task implementation, TDD workflow, or blueprint-driven development.

Creates structured handoff documents when a task needs different capability or has failed after retries. Use when flagging tasks for another session or model, after 2 failed fix attempts, or when architecture decisions are needed. Invoke for handoff, escalation, blocked task, failed attempts, or capability mismatch.

Project planning skill. Creates a comprehensive BLUEPRINT.md so any agent can implement without asking questions. Use when no BLUEPRINT.md exists yet and you need to plan a new project or feature. Invoke for project kickoff, planning, architecture design, or roadmap creation.

Communication style skill that enforces minimal, patch-only output. Always active. Governs how the agent formats responses -- prefer diffs over full files, commands over explanations, and brevity over verbosity. Invoke for output formatting, communication style, or response length control.

Self-modification skill for agent configuration files. Handles changes to AGENTS.md, skills, preferences, and state files with appropriate safety tiers. Use when the user asks to change agent rules, add skills, or update preferences. Invoke for rule changes, skill additions, preference updates, or APEX modifications.

Self-modification skill with tiered safety and rollback. Handles changes to AGENTS.md, skills, preferences, and state files. Triggers on phrases like "update rule", "add skill", "change how you", "remember that". Use when the user wants to modify agent behavior or configuration.