nerd

Plugin nerd in shawnroos/nerd. Commands read from the repo's manifest — 1 month ago.

Skills (7)

• codebase-analysis

  Reference for identifying sweepable experiment targets in codebases — tunable parameters and the values behind falsifiable questions. Use when scanning for research targets: hardcoded thresholds, magic numbers, heuristic weights, prompt templates, pipeline budgets.

• experiment-planning

  Reference for designing nerd experiments — competing theories, sweep harnesses, ground truth strategies, metric selection, and feasibility checks. Use when creating or reviewing experiment plans.

• intern-delegation

  Canonical delegation protocol for the nerd intern. Reference this when delegating tasks to the local LLM in /nerd or /nerd-this orchestrators. Defines health checks, timeouts, confidence gating, shadow comparison, fallback, and logging.

• intern-training

  Reference for intern training data formats, benchmark structure, and evaluation protocol. Use when running aptitude tests, collecting training data, or evaluating intern performance.

• performance-analysis

  Reference card for performance research — anti-pattern catalog, profiling tool reference, metric command templates, and measurability gate criteria. Use when writing performance experiment plans or analyzing performance findings.

• research-reporting

  Reference for compiling nerd findings into structured reports. Use when writing experiment reports, creating executive summaries, or documenting parameter tuning results.

• worktree-lifecycle

  Canonical, single-source procedures for nerd's experiment worktree lifecycle — create, merge-back, and reconcile/cleanup. Both /nerd and /nerd-this reference this so the safety logic lives in exactly one place instead of being copy-pasted (and drifting) across command files. Use whenever a nerd comm

Slash commands (7)

• /nerd:nerd-intern

  Train a local LLM to handle lightweight research tasks. Your intern starts dumb (really dumb) but learns from every nerd run and earns responsibility through demonstrated competence. Requires a local LLM serving stack (Ollama, MLX-LM, llama.cpp, or vLLM).

• /nerd:nerd-loop

  Run a continuous self-improvement loop on a specific aspect of your codebase. The agent edits code, runs it, measures the result, keeps improvements, discards regressions, and repeats indefinitely. Like Karpathy's autoresearch but for any codebase feature. Use: /nerd-loop 'search relevance' or /nerd

• /nerd:nerd-schedule

  Schedule nerd experiments to run at specific times (e.g., overnight). Uses macOS LaunchAgent for scheduling.

• /nerd:nerd-setup

  One-time global setup for the nerd plugin. Detects hardware, installs the training variant (MLX for Apple Silicon, original for NVIDIA), runs calibration benchmarks, and saves a hardware profile. Only needs to run once per machine — projects auto-initialize on first /nerd run.

• /nerd:nerd-status

  Check the status of the nerd queue, running experiments, and backlog. Shows progress, completed findings, and pending proposals.

• /nerd:nerd-this

  Context-scoped experiments. Researches only what you're working on right now — infers scope from your current branch, session files, and conversation topics, then groups findings into research themes and runs falsifiable experiments (numeric metric, or a pre-registered rubric judged by an LLM via ru

• /nerd:nerd

  Let the nerd loose on your codebase. Designs and runs rigorous experiments — any falsifiable question with a trusted instrument, whether a measurable numeric metric (parameter sweeps, single-commit hypothesis tests) or an LLM judge scoring against a pre-registered rubric — in worktrees while you sle

Subagents (10)

• context-scanner

  Scans a scoped set of files for tunable parameters and other sweepable experiment opportunities, and clusters results into research themes. Used by /nerd-this for context-scoped experiments.

• experiment-executor

  Executes nerd experiment plans in isolated worktrees. Builds evaluation harnesses and runs sweeps (parameter sweeps, single-commit comparisons, model/prompt A/Bs — any falsifiable experiment with a numeric metric), OR judges outputs against a pre-registered rubric (instrument: judge_rubric). Capture

• intern-evaluator

  Runs aptitude tests and ongoing evaluation of the local LLM intern. Calls the intern endpoint with benchmark examples for each task type (parameter-detection, result-classification, context-extraction), scores against expected outputs, and returns structured results with accuracy per task and mode r

• lab-tech

  Pre-flight validation agent that checks whether the lab is ready before experiments run. Verifies data access (WAL-mode, file permissions, exports), confirms config fields are actually wired in execution paths, scaffolds missing eval infrastructure (export scripts, test fixtures, datasets), and repo

• loop-scout

  Analyzes nerd research findings, experiment reports, and backlog proposals to identify the best candidates for deep /nerd-loop continuous improvement. Looks for areas with high improvement potential, measurable metrics, and clear scope boundaries. Use after /nerd completes or when deciding what to l

• parameter-scanner

  Scans codebases for tunable parameters, hardcoded thresholds, magic numbers, and empirical optimization opportunities — one source of falsifiable experiments nerd can run (others, like single-commit hypothesis tests, arrive as briefs rather than scans). Use when nerd needs to discover sweepable expe

• perf-explorer

  Maps codebase for performance research — identifies hot paths, I/O boundaries, complex functions, and areas of interest. Produces a structured area map that guides specialist agent dispatch.

• perf-specialist

  Parameterized performance specialist — analyzes a specific performance category (algorithmic, caching, io, memory, network) using anti-pattern catalogs from the performance-analysis skill. Dispatched by the orchestrator with a category parameter.

• plan-reviewer

  Reviews and improves nerd experiment plans. Generates competing theories, performs SpecFlow analysis, identifies gaps, and iterates until plans are robust. Use when experiment plans need quality review before execution.

• report-compiler

  Compiles nerd experiment results into structured research reports. Evaluates competing theories, identifies which held up and which were disproven, and delivers actionable recommendations. Use when experiments are complete and findings need to be documented.

Hooks (2)

• SessionStart
• PostToolUse

Similar plugins