nklisch

Goal-statement or direct-invocation queue driver for agile-workflow. Use when a harness goal or user request says to run autopilot, drain ready work, finish an epic, continue through .work/active/, or make autonomous progress on the substrate. Reads .work/active/, picks ready items by depends_on and stage, delegates to design, implement, and review skills, commits transitions, and repeats until the scope is done or blocked. Before reporting complete, runs a final peer-review/fresh-context completion pass and fixes or files accepted findings. No /loop or --resume mechanics; the harness goal/continuation feature owns long-running persistence. Epic-scoped by default; --all drains all active work; free-text scope directives are allowed.

Deep multi-angle correctness bug hunt. Use when the user asks to "scan for bugs", "bug hunt", "deep bug audit", "find hidden bugs", "find race conditions", or audit lurking correctness issues. Fans parallel scanner agents across relevant domains such as concurrency, async, state, resource leaks, time/numbers, error handling, data layer, and language footguns. Standalone mode writes a scored report; gate mode binds to a release bundle and emits .work/active/stories/ items with gate_origin:bugs. Distinct from gate-security, review, perf-design, and repo-eval.

Multi-lane, decomposition-first codebase scan campaign for agile-workflow. Use when the user asks for a deep code scan, comprehensive audit, whole-repo issue hunt, or organized fix plan that spans multiple lanes such as correctness, tests, performance, security, quality, structure, or architecture. Interviews for goal, lanes, rigor, scanner tier, and altitude bands; maps the repo from leaf to system scope; fans out scoped scanners; runs a review gauntlet; and consolidates surviving findings into one fix epic. For a single-domain scan, route to the specialist instead.

ALWAYS invoke this skill when the user asks to design e2e tests, audit or bootstrap the e2e program, or work on a [e2e-test]/[testing] feature at stage:drafting. Designs service-level-mocked e2e coverage, writes the design into the feature body, spawns child stories with depends_on chains, and advances drafting to implementing. Supports golden-path, failure-mode, chaos, and fuzzing coverage, plus --bootstrap and --audit modes.

ALWAYS invoke this skill when the user asks to design, decompose, or pick up an epic at stage:drafting. Reads the epic, foundation docs, and codebase; identifies child feature arcs; writes feature files with parent and depends_on metadata; updates the epic body with the realized decomposition; and advances drafting to implementing. When ux-ui-design is installed, includes mockup planning for net-new screens and journeys.

ALWAYS invoke when the user asks to design or flesh out a feature at stage:drafting; do not write design prose inline. Designs the feature inside its agile-workflow item body, grounded in foundation docs and code, then spawns child stories with depends_on chains and advances drafting to implementing. Use for greenfield features without [refactor], [perf], [prose], or [research] tags; route [refactor] to refactor-design, [perf] to perf-design, [prose] to prose-author, [research] to the agentic-research research-orchestrator, and epic decomposition to epic-design. UI/UX mockups are fallback here, inherited from the parent epic when available.

Cruft gate that scans the items bound to a release for AI-accumulated debris (dead code, stale comments, compatibility shims, defensive bloat, over-abstraction) introduced or revealed by the bundle. Delegates the full scan to a deep cleanup scanner agent which runs language-aware detection plus heuristic pattern-matching, then returns findings. The orchestrator converts findings into items in .work/active/ with gate_origin:cruft and tags:[cleanup]. Auto-triggers during /agile-workflow:release-deploy.

Documentation gate that enforces the rolling-foundation principle. Delegates the full drift detection to a deep documentation scanner agent which scans the bundle's changes for foundation-doc drift (assertions in docs/ that no longer match implementation), changelog gaps, README staleness, and skill/pattern-skill staleness. The orchestrator converts findings into items in .work/active/ with gate_origin:docs and tags:[documentation]. Auto-triggers during /agile-workflow:release-deploy.

Project code patterns and conventions. Auto-loads when implementing, designing, verifying, or reviewing code. Provides detailed pattern definitions with code examples. Consult this whenever writing new code, reviewing changes, or trying to understand how this project structures things.

Code review skill from tap plugin

A helper skill

Lints code

Research findings on Claude Code marketplace, plugin structure, and distribution. Auto-loads when working with marketplace.json, plugin.json, plugin distribution, Claude Code plugin format, SKILL.md frontmatter, tap.json integration, npm skill packages, agentskills.io standard, /plugin install, or designing how skilltap interacts with the Claude Code plugin ecosystem.

Biome v2 linter/formatter reference. Auto-loads when working with biome.json, biome config, linting rules, formatting, import organization, biome check.

Zod v4 validation library reference. Auto-loads when working with Zod schemas, validation, z.object, z.string, z.infer, safeParse, transforms, discriminatedUnion.

Reference for Bun runtime APIs used in this project. Use this skill whenever writing code that runs shell commands (Bun.$ or Bun.spawn), reading/writing files (Bun.file/Bun.write), running tests (bun test), or configuring the monorepo (workspaces, bunfig.toml). Prefer Bun APIs over Node.js equivalents — use Bun.$ instead of child_process, Bun.file() instead of fs.readFile, bun:test instead of jest.

Browser recording and control via the krometrail CLI. Use when recording a browser session, driving the browser with batch actions, or investigating recorded sessions. Captures network, console, DOM, framework state, screenshots.

Krometrail MCP navigation guide. Load this skill first to understand which tool namespace to use. Covers the three tool namespaces (debug_*, chrome_*, session_*), when to use each, common pitfalls, and pointers to the detailed skill for each namespace.

Runtime debugging via the krometrail CLI. Use when a test fails and reading the code isn't enough, when you need to inspect runtime values, or when a bug is in logic you can't trace statically. Set breakpoints, step through code, inspect variables, evaluate expressions — across 10 languages.

Project structural organization rules for TypeScript/Bun. Proactively scans for organizational issues and produces a prioritized plan. Defines the team's preferred file, folder, and module structure.

Project stylistic refactoring rules for TypeScript/Bun. Proactively scans for refactoring opportunities and produces a prioritized plan. Defines the team's preferred coding style.

Project code patterns and conventions. Auto-loads when implementing, designing, verifying, or reviewing code. Provides detailed pattern definitions with code examples.

React DevTools hook internals reference. Use when implementing or debugging React state observation — fiber tree traversal, hook linked lists, onCommitFiberRoot, state extraction, bug pattern detection.

SolidJS debug interface reference. Use when implementing or debugging Solid state observation — DEV hooks, signal tracking, ownership tree, store observation. Tier 2: requires dev-mode builds.

Godot scene and resource authoring via Director MCP tools or CLI. ACTIVATE when: user asks to create scenes, add/remove/reparent nodes, set node properties, create materials or shapes, edit tilemaps or gridmaps, create or modify animations, connect/disconnect signals, set collision layers, attach scripts, diff scenes, manage autoloads, set project settings, check for script errors, see editor state, or perform any Godot project file editing task programmatically. Also activate for batch operations or "build me a level/scene/UI". Do NOT activate for observing a running game — use theatre-stage for that.

Project code patterns and conventions. Auto-loads when implementing, designing, verifying, or reviewing code. Provides detailed pattern definitions with code examples.

Research findings on the schemars Rust crate for JSON Schema generation. Auto-loads when working with schemars, JsonSchema derive macro, json_schema! macro, schema_for!, SchemaSettings, SchemaGenerator, JSON Schema from Rust types, schemars attributes, schemars feature flags, or migrating from schemars 0.8 to 1.x.

Spatial debugging for running Godot games via Stage MCP tools or CLI. ACTIVATE when: user mentions game state, node positions, spatial bugs, physics issues, pathfinding problems, collision debugging, AI behavior, signal tracing, "take a snapshot", "what's happening in the game", clip/recording analysis, watch subscriptions, or any task that requires observing or interacting with a live Godot game world. Also activate for frame-by-frame debugging, teleporting nodes, pausing/advancing time, or injecting input. Do NOT activate for editing .tscn files or creating scenes — use theatre-director for that.

Working with godot-rust/gdext (Rust GDExtension bindings for Godot) in the stage-godot crate. Use when writing or modifying Rust code that compiles to the GDExtension library.

Working with the Godot addon GDScript layer in addons/stage/. Covers EditorPlugin patterns, autoload management, dock panels, input handling, and GDExtension integration from the GDScript side.

Working with the rmcp Rust MCP SDK in the stage-server crate. Use when writing or modifying MCP tool definitions, server initialization, or tool call handling.

Orientation for working on the Stage codebase itself. Covers repo layout, crate responsibilities, design decisions, key patterns, and how the pieces connect.

Use when debugging Praxis agent-harness failures: tool calls leaked into chat, raw <invoke> markup, [object Object] rendering, tool.dispatch.error, sub-agent stalls, IPC stream failures, React renderer crashes, persistence/FK failures, course-create drafting failures, or student-simulation/browser mismatches. Guides agents through local evidence capture, replay/simulation commands, owner routing, and substrate item filing.

Project code patterns and conventions. Auto-loads when implementing, designing, verifying, or reviewing code. Provides detailed pattern definitions with code examples.

Project-specific refactor conventions for Praxis. Auto-loads when agile-workflow refactor-design scans or designs refactors. Extends refactor-design's default smell scan with team-confirmed style and structural preferences; does not replace the defaults and does not create standalone plan docs.

Reference for the Gemini / Antigravity engine path in Praxis. Auto-loads when an agent is touching Gemini, Antigravity, `@ai-sdk/google`, `direct.google`, `google-antigravity`, `GEMINI_API_KEY`, `GOOGLE_GENERATIVE_AI_API_KEY`, or considering a new engine adapter for Google models. Captures the policy + technical constraints found in `docs/research/gemini-subscription-engine.md` (2026-05-19) so future agents don't re-discover them.

officeparser v6 reference for parsing .pptx (and other office formats) into a typed AST with extracted attachments. Triggers on imports from `officeparser`, mentions of `OfficeParser.parseOffice`, `OfficeParserAST`, `OfficeAttachment`, `OfficeContentNode`, or work on PowerPoint / PPTX ingestion in `packages/tools/src/runtime/ingestion/`.

Drizzle ORM v0.45+ reference for PostgreSQL. Auto-loads when working with drizzle-orm, pgTable, schema definitions, drizzle queries, migrations, drizzle-kit.

Zustand v5 state management reference. Auto-loads when working with zustand stores, create, createStore, useShallow, persist, devtools, immer middleware.

Compare project base prompts and model metadata against a new Claude Code release. Extracts upstream system prompt, diffs each fragment, classifies changes as intentional omission or drift, and applies approved updates. Use when user says "sync upstream", "check for drift", "new CC version", or runs /sync-upstream.

Project structural organization rules for Bun/TypeScript CLI. Proactively scans for organizational issues and produces a prioritized plan. Defines file, folder, and module structure preferences.

Project code patterns and conventions. Auto-loads when implementing, designing, verifying, or reviewing code. Provides detailed pattern definitions with code examples.

Project stylistic refactoring rules for Bun/TypeScript CLI. Proactively scans for refactoring opportunities and produces a prioritized plan. Defines the team's preferred coding style.

Run iterative cross-model peer review on substantive designs, code, plans, refactors, architecture, docs, or repo artifacts by delegating to the other local coding agent through peeragent. Use when the user asks for peer review, a second opinion, cross-review, feedback from Codex/Claude/Gemini/GLM, or when recent work is risky enough to deserve an independent review loop. Resolve the bundled wrapper from the plugin location; do not assume peeragent is on PATH.

Delegate arbitrary task work — implementation, research, review, debugging, design, doc updates, anything — to another local coding agent through the bundled peeragent wrapper. Use when the host assistant wants a peer agent to take a focused pass on a task in the current repository. Targets: Codex (default), Claude Code, Gemini through Antigravity, or Z.AI GLM 5.2 through Pi. Free-form task text; the wrapper returns a JSON result the host summarizes for the user. Resolve the wrapper from the plugin location before invoking; do not assume `peeragent` is on PATH.