A fast, low-friction execution skill that bypasses heavy planning phases to immediately patch linter errors, syntax issues, or stack traces. Triggered by saying "/auto-fix" or pasting an error log. It finds the file, applies the minimal fix, and stops for…
Run any question, idea, or decision through a council of 5 AI advisors who independently analyze it, peer-review each other anonymously, and synthesize a final verdict. Based on Karpathy's LLM Council methodology. MANDATORY TRIGGERS: 'council this', 'run the…
Enforces a strict planning phase before execution. For complex tasks, it triggers a 'grill-me' style Q&A and generates an implementation plan using Ponytail discipline (lazy-senior principles) and YAGNI. For simple tasks, it skips heavy planning.
Operates in strict advisory and consultant mode (read-only safety guardrail + code and plan review). Prevents file modifications, code edits, or destructive commands. Evaluates code diffs, architecture, and execution plans using Ponytail discipline…
Interactively grills the user with technical and architectural questions to uncover missing requirements, design constraints, and edge cases before implementation. Trigger when the user types `/grill-me`, "grill me", or requests interactive requirement…
Suggests or executes a git commit based ONLY on currently staged files. Warns and stops immediately if no files are staged. Supports simple mode (title-only), detailed mode, and auto-commit mode (`git commit -m <title> [-m <desc>]`). Learns repository commit…
Takes over the user's active tmux terminal session (default matching pattern `wasd-*`). Checks active tmux sessions without a pager, prompts the user if multiple sessions exist, and remembers the session name in memory context.
Instructs the agent to output requested text, code snippets, or markdown directly into the chat interface instead of creating or modifying files. Trigger when the user explicitly asks to "print", "show", or output something in the chat.