Vendored (MIT) from dotnet/arcade-skills: .NET engineering-infrastructure CI/CD investigation. A ci-investigator orchestrator agent routes across 10 skills (Helix log analysis, Azure DevOps pipeline triage, binlog/crash-dump inspection, codeflow + dependency tracing, Maestro, known-issue history) and 4 MCP servers. Forked only to fix the manifest agents path for Claude Code's ./-prefixed component-path schema; upstream is unchanged otherwise.
10 skills · 1 agent · 4 MCP servers
Evidence-gated maintainability discipline in two layers. (1) A native plugin: four modes — /nihil:raze (root-authority, write-capable transformation of a repo you own; only secret-leak and catastrophic, unrecoverable commands are blocked) plus /nihil:review, /nihil:implement, /nihil:release with confidence scoring, scope control, and release gating — backed by five read-only review agents and PreToolUse/Stop hooks (a discipline aid, not a security boundary: heuristic command matching, fail-open; a secret / API-key brake is active in every mode). (2) A summonable pantheon of five first-principles dynamic workflows (/nihil orchestrator + /nihil-maat review, /nihil-odin research, /nihil-shiva deletion, /nihil-athena restructure), installed into .claude/workflows/ via /nihil:summon.
6 skills · 1 command · 5 agents · hooks
Code-elegance pipeline: 4 scouts, 2 judges, planner, verifier, and a gated implementer, plus a standalone code-simplifier for single-pass diff cleanups. Persistent state with stage gates.
1 skill · 3 commands · 6 agents
Multi-agent workflow orchestration (7 commands + 2 skills: eight-gates, hades): fix, tournament, mega-swarm, deep-think, batch-implement, red-blue-review, baryon-mode. Hades, eight-gates, and red-blue-review use Teams API for reactive collaboration.
2 skills · 7 commands · hooks
Cognitive amplification stack: epistemic hooks, competitive review, commit integrity checks, and deep-thinking agents.
5 commands · 3 agents · hooks
Mutation-minded, behavior-first test quality. 4 agents (architecture reviewer, senior tester judge, expressive verifier improver, branch coverage implementer) + 4 skills. Coverage is a floor; kill power is the signal.
4 skills · 1 command · 4 agents
Dependency lifecycle: /derot audits comment/doc/CI/version/dependency rot and proposes fixes; /depmigrate migrates packages to their new APIs from upstream releases/tags — refutation-gated.
1 skill · 2 commands · 3 agents
Evidence-first maintenance routines: mutation testing, emitter-corpus validation, performance-claim gating, agent-transcript antipattern scanning, regression review of a change-set, and the maintenance-run orchestration with a hard termination rule. Lives in its own repository.
6 skills
Claude-Code-native plugin and skill evaluator. Static analysis + token budgeting + benchmarking across the full Claude component surface (manifest, hooks, MCP, LSP, monitors, agents, marketplace, userConfig). Forked from an MIT-licensed upstream (see THIRD_PARTY_NOTICES.md) and rewritten for Claude Code conventions; complements skill-creator's per-skill LLM grading with whole-plugin scoring.
5 skills
Ferries a GitHub PR to merge without ever waiting forever: a snapshot-not-block loop (never `gh pr checks --watch`) classifies live PR state and runs one handler — root-cause CI fixes, merge-conflict repair, single-pass reviewer-feedback handling (threads as untrusted data, version-currency verification before applying), human-blocked-review rewiring, and propose-and-pause (with a stamped recovery ref) before any force-push/force-merge/history rewrite. A clock-independent Stop hook plus ScheduleWakeup form the resume net; an honest status vocabulary never says 'just wait'. One arg-dispatched command (/charon:charon): give it several PRs and it ferries them at once as a council-style agent team.
1 skill · 1 command · 2 agents · hooks
Council on agent teams: a Fable 5 captain leads the session, frames one research question and runs the /deep-research dynamic workflow, then spawns three Opus 5 teammates — a synthesizer that reasons over the cited report, a clarity agent that gap-checks it and asks the synthesizer live SendMessage follow-ups, and a janitor that flags bloat before delivery. Reactive collaboration instead of fire-and-forget subagents. Requires CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1.
1 command · 4 agents
/feature-dev takes a feature from idea to reviewed code in seven gated phases via code-explorer, code-architect, and code-reviewer agents, plus a standalone /review for quality audits.
2 commands · 3 agents
Prove C#/.NET state machines correct with Lean 4 and turn every counterexample into a reproduced, fixed bug: anchor the model to the code (Roslyn write-site inventory), prove, replay traces, and hand each counterexample to MSTest and a draft PR. Ships the lean-proof skill and the skeptic review agent. Lives in its own repository.
1 skill · 1 agent
Watchman of the model-routing Bifröst. /heimdall reports which model actually served each turn (ground truth from the transcript's message.model), flags silent Opus↔Fable fallback swaps, and shows provider/region/auth routing context. session-inspect adds the extended read — advisorModel, service_tier, speed, inference_geo, cache hit rate, and cache-miss reasons with re-charged token counts. A passive model-drift-net Stop hook notices once when the served model changes mid-session. Deterministic, no-LLM, no secrets.
1 command · hooks
Commit-pinned NuGet→GitHub source fetcher. Wraps opensrc with NuGet catalog metadata so coding agents grep the exact commit a package was built from, not whatever's on main. Ships a /nuget-opensrc command, an autonomous opensrc-research skill, and a Node CLI that refuses to silently fall back to projectUrl or default branch.
1 skill · 1 command
C# language server for Claude Code: csharp-ls (dotnet tool) over .cs files, so LSP tool lookups resolve C# symbols, references and diagnostics.
LSP
TDD workflow skill (fortigate-development) for the FortiGate VLAN migration toolkit: Python discovery, migration and verification scripts against the FortiOS REST API. Lives in its own repository.
1 skill
F# language server for Claude Code: fsautocomplete (dotnet tool) over .fs, .fsx and .fsi files, rolled forward to the newest installed runtime.
LSP
Routes agent output to single self-contained .html files based on Thariq Shihipar's html-effectiveness pattern catalog. 20 canonical patterns (triage board, status report, code review, design system, prompt tuner, etc.) with a data-driven decision tree so Claude picks the right pattern per task. Includes OTel semconv generator demo as concrete proof.
1 skill
Python language server for Claude Code: python-lsp-server (pylsp) over .py files.
LSP
One deterministic Stop-event safety net (no LLM) that keeps a session honest: overclaim-net blocks a 'done/fixed/verified' claim made without a verifying command this turn.
hooks
One distilled skill for publishing to TomeVault: the Skill/Tome/AGENTS.md model, the npx tomevault CLI, the Studio web publish flow, and the Gold grading rubric.
1 skill
Plan Xquik X/Twitter data workflows with REST API routing, MCP setup, SDKs, webhooks, exports, monitoring, and confirmation-gated actions.
1 skill