dotnet-reference

Batch-verify and delete dead code: rg-confirm zero callers for N symbols, delete files, run the build once, stage results.

Run benchmarks and check regressions, add benchmark gating to CI, or audit GitHub Actions workflows. Determines mode from user intent.

Batch-apply findings from the most recent --local review. One confirmation, one test run.

Orchestrated code review of git diff using parallel agents. Standard mode (diff-scoped) or full mode (adds Roslyn semantic analysis). Produces severity-grouped plan.

Unified code review: --local (diff review), --pr (adversarial PR review), --apply (batch fixes from last review).

Audit .NET solution health across 13 dimensions: dependency graph, solution sync, CPM compliance, package health, framework consistency, IVT & encapsulation, build config, analyzer config, NuGet config, suppression hygiene, output type & AOT, public API surface, and lock files & pruning. Works with any .NET solution (.slnx or .sln). Parses all config files once, spawns 3 parallel agents, and produces a severity-grouped report with a Mermaid dependency graph. Triggers on: solution audit, audit solution, project graph, dependency audit, solution health, audit projects, check solution.

Scans source and .claude/ files for anti-pattern violations, applies the top mechanical fixes, and reports pre/post counts.

Recursively analyze the codebase and generate nested Mermaid architecture diagrams with structured annotations. Includes dependency mapping (NuGet packages, infrastructure, runtime pinning) for .NET projects.

Write, review, and fix Bash (.sh) scripts using project conventions and best practices. Always use this skill when writing any .sh file, creating a new shell script, fixing a bash error or runtime error, reviewing shell code, adding a new script to scripts/, or when the user asks about bash style, naming, formatting, or patterns. This skill combines official style rules with project-specific patterns to produce correct, portable scripts on the first try. Trigger on: write bash, create .sh, bash script, shell script, bash error, fix .sh, bash style, bash patterns, script error in bash.

Code coverage skills: fix low coverage, enforce thresholds in CI, or audit coverage.config.xml for drift. Determines mode from user intent automatically. Triggers on: coverage fix, low coverage, improve coverage, fix low coverage, coverage gate, coverage threshold, enforce coverage, minimum coverage, coverage enforcement, add coverage gate, audit coverage config, coverage config drift, coverage.config.xml, check coverage targets, fix coverage config, stale coverage config, coverage missing project, coverage work, coverage skill.

Generates or updates a lode dependency map for a scope. Scans all repos under a path for NuGet, npm/yarn, Maven, Go, Rust, and Python dependencies; detects infrastructure from config files; writes lode/<scope>/dep-map.md with per-repo Mermaid service-layer diagrams and a cross-repo dependency heat map.

Opinionated .NET code overhaul loop. Establishes a test baseline, modernizes language usage, fixes cross-cutting design issues, runs performance, concurrency, and security reviews, and verifies all CI tests pass. Presents severity-rated findings for user-directed action at each step. Review conventions.md before first use — preferences are configurable. Use iteratively for initial large overhauls, then periodically to catch drift. Triggers on: code overhaul, modernize codebase, .NET modernization, overhaul loop, overhaul pass, fix all warnings, language modernization, primary constructors, collection expressions, overhaul solution.

Investigate failing tests or CI divergence. Deep single-test root-cause analysis, or local-vs-CI pattern matching.

Fix build errors in batch. Parses pasted CI/build output, classifies errors, applies minimal fixes, validates. For test investigation or CI divergence, use /fix-investigate instead.

Roger Martin strategic alignment audit for code investment decisions.

Migrate C# types between namespaces across the solution. Updates usings, refs, and verifies build.

Manage C# naming rules and bulk-rename violations. --rule edits .editorconfig, --migrate bulk-renames IDE1006 violations.

Build a pyramid-structured wiki for a product area by exploring its source repos. Produces Tufte Dracula HTML docs with Mermaid diagrams.

Orchestrates all code review agents to audit recent .NET changes. Accepts --commits N or --days N to define review scope, discovers agents dynamically, runs them in parallel, and produces a severity-grouped report with actionable recommendations. Triggers on: code review, review code, review changes, audit code, check code quality, clanker check, state bloat, grab bag review, check for flag bloat, mutation review, clanker discipline.

Find repeated behaviors in Claude Code sessions that should be deterministic bash scripts. A "deterministic script" replaces an ad-hoc multi-turn Claude behavior with a repeatable, token-free automation (bash script + optional hook + optional keyword shortcut). Scores candidates on token-savings potential and builds the top N if approved. Supports --scan-skills mode to statically analyze agent/skill files for inline bash that should be extracted to scripts. Triggers on: script discovery, find script candidates, what should be a script, repeated patterns, deterministic scripts, automate patterns, session patterns, discover scripts, what keeps repeating, find automations, script opportunities, review agents for scripts, extract scripts from skills, deterministic portions. Not for: token optimization plans (use token-optimizer), weekly session analysis (use weekly-learner), or single-session review (use token-optimizer --current).

Tech debt and architecture audit for .NET repos. Produces plans/TECH_DEBT_AUDIT.md with file-cited findings, severity, and effort estimates. Supports module scoping and focused-mode shortcuts (crap, dead-code, duplicates, suppressions, solid) for single-dimension audits. Does not auto-invoke. Triggers on: tech debt audit, debt audit, codebase health check, architecture review, code quality assessment, audit tech debt, tech debt scan, code health, crap analysis, crap score, dead code audit, find dead code, unused types, dup-scan, code duplication, duplicate code, find duplicates, DRY violations, review code suppressions, audit pragma warnings, pragma warning disable, suppression cleanup, code hygiene, SOLID violations, solid review.

STRIDE threat modeling with CAPEC drill-down and DREAD-lite scoring. Builds a DFD from lode/ domain knowledge, identifies threats per trust boundary, maps to concrete attack patterns, and produces a prioritized threat register.

Ship it — formats, builds, tests, commits, and pushes. Accepts --skip-tests flag to skip build+test when caller already verified them. Triggers on: ship it, yeet, push it, commit and push, deploy this, we're done, preflight, preflight check, quality check, pre-commit check, ready to commit.

Periodic health check for the entire Claude Code configuration — agents, skills, CLAUDE.md, and lode memory. Optionally fetches latest Anthropic guidance (web research is skippable with --no-web), reviews all agents for model/tool/memory optimization, reviews all skills for token efficiency, audits CLAUDE.md files, checks agent memory health and git tracking, and writes a dated revision log to lode/ for cross-session continuity. Run monthly or after bulk agent/skill changes. Triggers on: claude revision, periodic audit, agent health check, skill optimization, claude config review, monthly claude audit, revision audit, config health, claude maintenance, agents audit, skills audit, claude audit.

Strategic alignment audit for codebases. Applies Roger Martin's strategy framework directly to code — analyzing what the code reveals about its strategic choices, investment levels, and capability reinforcement. Produces a prioritized findings report and feeds actionable refactorings into dev-planning. Use for code review through a strategy lens, feature gap analysis, investment mismatch detection, or surfacing undocumented strategic choices. Triggers on: strategy audit, strategic audit, martinize, martinizing, roger martin, code as strategy, strategy review, investment mismatch, capability analysis, strategic alignment, where should I invest, code investment audit, feature gap analysis, over-engineered, dead code strategy, what's table stakes.

Find and remove unused NuGet packages and stale project references in a .NET solution. Reviews Directory.Packages.props for PackageVersion entries no project references, reviews every .csproj for PackageReference and ProjectReference entries the project's code no longer uses, then removes approved entries and verifies the build. Avoids the false-positive traps (transitive pins, analyzers, runtime-only packages). Triggers on: unused packages, prune dependencies, dead dependencies, remove unused packages, unused package references, unused project references, stale references, clean up Directory.Packages.props, unneeded references, dependency prune, orphaned packages, are these packages still used.

Weekly error triage skill. Runs scripts/session-health.sh --json to get error counts by category, surfaces the top-3 recurring patterns with ranked fix suggestions, and identifies actionable root causes (flag errors, write-before-read, parallel cascade failures, hook-denied patterns, etc.). When permission errors are in the top 3, runs a settings.json gap analysis inline (formerly allowed-tools-maintainer). Triggers on: error audit, what's breaking, error report, session health, what errors keep happening, why do I keep getting errors, recurring errors, tool errors this week, fix recurring errors, settings sync, allowed tools gap, tool approval friction, settings audit, add to allow-list, skills need permissions, sync tool permissions, settings.json gaps. Not for: fixing a specific one-off error (use /fix-ci).

Audits all README.md files in the repo for staleness after significant code changes and auto-fixes drifted tables. Discovers READMEs dynamically. Skips automatically when fewer than 50 C# files changed in recent history. Triggers on: check readme, readme stale, is readme up to date, readme check, readme accuracy, readme outdated, update readme.

Researches a topic from the web and adds it to the project's lode/ documentation. Use /learn <topic> to research official docs and persist findings. Supports --skill flag to output as a SKILL.md instead. Triggers on: learn about, research topic, add to lode from web, learn this, deep research.

End-to-end investigation of a single failing unit test. Accepts a test name or pasted xUnit output, runs the test in isolation, reads the test and source files, categorizes the failure, and proposes a targeted fix. Distinct from fix-ci which handles batch CI output; this skill goes deep on one test at a time. Triggers on: debug this test, investigate failing test, why is this test failing, fix failing test, test debugging, debug test, investigate test failure, test fails, single test debug, test broken, why does this test fail.

Capture a single insight into the lode/ documentation store. Routes the insight to the correct lode file based on topic, writes it as a current-state fact, and timestamps the file. Lightweight alternative to lode-sync for one-off captures. Triggers on: save this to the lode, capture this insight, capture to lode, remember this for next time, save this knowledge, capture that knowledge.

Audits lode/ documentation for content accuracy — checks whether each doc correctly describes the current codebase. Distinct from lode-sync (which only checks timestamps). Produces a staleness table with git-verified findings, then makes targeted updates only on the sections you approve. Use after large refactors, renames, or when the lode feels out of sync with the code. Triggers on: audit lode, lode out of date, lode accuracy check, is the lode correct, lode stale, lode content review, audit documentation, check lode, lode drift, lode wrong, lode inaccurate, lode needs correction. Not for: updating a single known lode file (just edit it directly), or checking timestamps only (use lode-sync agent).

Quick lookup in the lode/ documentation store. Searches for a topic across all lode files and returns matching content with context. Fast alternative to manually grepping the documentation tree. Triggers on: lode search, search lode, find in lode, lode lookup, what does lode say about, check lode for, lode docs.