pbdeuchler

pbdeuchler/llm-plugins

Plugin Claude Code

Collection of plugins for claude code.

pbdeuchler/llm-plugins

Instructions file

Claude Code instructions for pbdeuchler/llm-plugins, covering llm-plugins, conventions, task invocations use xml syntax and version updates require marketplace and changelog sync.

pbdeuchler/llm-plugins

Plugin Claude Code

Autonomous experiment loop for optimizing any measurable target.

pbdeuchler/llm-plugins

Agent

Use when executing a single autoresearch experiment iteration - implements code changes, runs benchmark, evaluates metrics, logs results to JSONL, and manages git state (commit or revert). Returns structured result to the orchestrator.

pbdeuchler/llm-plugins

Command

Start or resume an autonomous experiment loop with optional time limit.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Set up and run an autonomous experiment loop for any optimization target. Gathers what to optimize, then starts the loop immediately. Use when asked to "run autoresearch", "optimize X in a loop", "set up autoresearch for X", or "start experiments".

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Compute and interpret MAD-based confidence scores for experiment results. Use when logging experiment results after 3+ data points to determine if improvements are real or within noise.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Git commit and revert patterns for autoresearch experiments. Use when keeping or discarding experiment results to manage git state correctly.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Parse METRIC output lines, infer units, and track primary vs secondary metrics. Use when processing experiment output from autoresearch.sh.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Manage autoresearch.jsonl logging, session initialization, segment tracking, and session recovery. Use when starting, resuming, or recording experiments.

pbdeuchler/llm-plugins

Plugin Claude Code

Multi-perspective codebase review with severity-classified findings and holistic remediation.

pbdeuchler/llm-plugins

Agent

Use when performing a structured codebase review from a blank slate -- no existing plan or prior context. Receives a structural inventory and sampled file contents, analyzes across seven dimensions, and returns severity-classified findings with holistic remediation. Opus-powered multi-persona staff engineer panel.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Use when performing a blank-slate codebase review -- orchestrates the scout, file sampling, fan-out decision, reviewer dispatch, and output formatting pipeline.

pbdeuchler/llm-plugins

Plugin Claude Code

Deterministic guardrails and skills that prevent recurring user-correction patterns. Sister plugin to house-style.

pbdeuchler/llm-plugins

Hook

Runs after a tool call finishes for Write and Edit tool calls, executing flag-pattern-comments.py via python3. From pbdeuchler/llm-plugins.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Use when the user provides a reference implementation (URL, repo path, branch, file) plus a verb like "use", "copy", "adapt", "port", "model after" - mandates fetching/reading the reference before designing, and lifting the structure rather than reimagining it.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Use when adding or modifying non-trivial code - requires happy-path AND sad-path tests covering every return permutation, blocks declaring "done" if production code grew without matching test additions.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Use when editing rendered UI, dashboard, chart, or any visual artifact (Grafana, Terraform-generated panels, frontend components, plot output) - blocks declaring "done" without an external verification artifact.

pbdeuchler/llm-plugins

Plugin Claude Code

House style for software development.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Use when asked to write, draft, file, or open a GitHub issue - mandates exploring the relevant code to understand the whole problem space, reasoning any proposed solution, and structuring the issue with summary, what/why, references, nuances, open questions, and acceptance criteria. Adds professionalism and…

pbdeuchler/llm-plugins

Skill Claude CodeCodex

ALWAYS use before any implementation work, review, or planning: designing, writing, editing, refactoring, debugging, reviewing, or generating source code, tests, scripts, config-as-code, or behavior-affecting files. This is a baseline engineering skill and should be invoked even when a more specific language…

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Use when invalid data causes failures deep in execution - validates at every layer data passes through to make bugs structurally impossible rather than temporarily fixed.

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Use when reviewing a completed implementation holistically - a finished branch, a completed plan step, or a diff before PR. Reviews every changed line against the implementation plan (if one exists) and high engineering standards across seven dimensions, verifies tests by deliberately breaking crucial code, and emits…

pbdeuchler/llm-plugins

Skill Claude CodeCodex

Use when writing database access code, creating schemas, or managing transactions with PostgreSQL - enforces transaction safety with TX naming, read-write separation, type safety for UUIDs/JSONB, and snakecase conventions to prevent data corruption and type errors.