claude

kstack

k-mode router + playbooks, per-role Haiku/Sonnet/Opus effort tiers, k-effort level switch, k-tokens audit and token settings, k-spectre research, k-war-room expert panels, k-review, k-e2e-qa, and safety hooks.

17 on GitHubby Kasempiternalv1.4.0MITupdated 15 days ago
/

Plugin kstack from marketplace k-stack in Kasempiternal/K-stack. Commands read from the repo's manifest — 13 days ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use. The exact commands are listed below; read them before installing.

What it installs

Skills (14)

  • k-agent-time

    Use when asked to improve, optimize, redesign or solve an open-ended technical problem, or before dismissing an option as too big, slow or costly. Prices effort for agents, not human-hours.

  • k-cursor-architect

    Delegate focused implementation or investigation from Claude Code to the Cursor Agent CLI (cursor-agent) running Grok 4.7. Use when the user invokes /k-cursor-architect, asks Claude to hand work to Cursor or Grok, or enables session-only delegation. Session mode makes Cursor the exclusive delegation

  • k-design

    Design before implementing. Opus produces the data shape, interface, file plan, failure cases, tests, and verification commands for a change, and runs a competing-designs bakeoff when the design is contested or has no precedent. Use for architecture, system or API design, implementation plans, repo

  • k-e2e-qa

    E2E and UI verification workflows. AgentController for native macOS apps (mcp__agentcontroller__*) and for iOS simulators and physical iPhones (mcp__agentcontroller-ios__*), Maestro or raw adb for Android (mcp__maestro__*), Claude in Chrome or Playwright MCP for the web. Use when running UI tests, Q

  • k-effort

    Switch K-stack's routing level the way /effort switches session effort. Levels are lean < balanced < quality, plus burn on top. A typed level applies now and is saved as the default; add "session" to keep it to this session. Use for /k-effort, "kstack effort", "cheaper kstack routing", or "more effo

  • k-explore

    Explain how a subsystem works, trace a runtime flow, answer where code should live, or dig up why something was built this way (git history). Read-only. Use for "how does X work", "walk me through", "why is this like this", "where should this live", onboarding onto unfamiliar code, or from k-mode be

  • k-gpt-architect

    Delegate focused implementation or investigation from Claude Code to the Codex MCP. Use when the user invokes /k-gpt-architect, asks Claude to hand work to Codex, or enables session-only delegation. Session mode makes Codex the exclusive delegation backend and blocks native Claude teammates. Never e

  • k-mode

    Kasempiternal's Claude-only engineering mode. The lead plans and reviews, work is routed by role to Sonnet or Opus effort tiers, every task follows a playbook, and nothing is called done without evidence from the real artifact. Use for /k-mode or requests to work in K-stack style.

  • k-review

    Review a diff with Opus reviewers on separate lenses (correctness, simplification, security and data, tests), judge the findings as a pragmatic lead, and route fixes to Sonnet in one batch. Use for code review, "review this", "tear this apart", "find blind spots", a final pre-release sweep, or from

  • k-setup

    Configure K-stack's role-to-tier routing (which Haiku, Sonnet or Opus effort each role uses). Writes a managed kstack-routing block into ~/.claude/CLAUDE.md so routing loads in every session. Use for /k-setup, "configure kstack", "kstack preset", or changing K-stack's model routing. For a quick leve

  • k-spectre

    Parallel research with synthesis and independent claim validation. Use for multi-source research questions, answers that need current external evidence, comparing tools or approaches, and questions that span the codebase and the web. Not for single-fact lookups. Invoked as /k-spectre <research quest

  • k-tokens

    Cut Claude Code token use. Audit reads your local transcripts and shows where tokens go (idle cache rewrites, context size, polling, Opus agents). Apply is a one-screen settings choice that merges a few documented keys into ~/.claude/settings.json and adds agent working rules to ~/.claude/CLAUDE.md.

  • k-war-room

    Convene a moderated war room. A panel of expert personas played by Claude models (mixed tiers; external models only on request) studies a task, opens with grounded positions, debates live in a Slack-style @mention chat the user watches in the browser, votes, and writes a transcript plus a verdict (d

  • k-write

    K-stack prose rules for docs, READMEs, RFCs, PR descriptions, commit bodies, code comments, and agent replies. Cuts AI tells and routes writing to the right effort tier. Use when writing or editing any prose, or from k-mode.

Subagents (11)

  • k-haiku-high

    K-stack tier worker: Haiku 5.5 at high effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-haiku-medium

    K-stack tier worker: Haiku 5.5 at medium effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-haiku-xhigh

    K-stack tier worker: Haiku 5.5 at xhigh effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-opus-high

    K-stack tier worker: Opus 5.5 at high effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-opus-max

    K-stack tier worker: Opus 5.5 at max effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-opus-medium

    K-stack tier worker: Opus 5.5 at medium effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-opus-xhigh

    K-stack tier worker: Opus 5.5 at xhigh effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-sonnet-high

    K-stack tier worker: Sonnet 5.5 at high effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-sonnet-low

    K-stack tier worker: Sonnet 5.5 at low effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-sonnet-medium

    K-stack tier worker: Sonnet 5.5 at medium effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

  • k-sonnet-xhigh

    K-stack tier worker: Sonnet 5.5 at xhigh effort. Dispatched by k-mode roles; see k-mode/references/routing.md. Not for direct invocation.

Hooks (9)

  • SessionStart

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/notify-session-start.sh"

  • SessionStart

    node "${CLAUDE_PLUGIN_ROOT}/skills/k-effort/scripts/k-effort.js" hook

  • Notification

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/notify-attention.sh"

  • Notification

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/notify-attention.sh"

  • Stop

    bash "${CLAUDE_PLUGIN_ROOT}/hooks/notify-stop.sh"

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/hooks/protect-secrets.js"

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/hooks/sensitive-command-approval.cjs"

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/hooks/no-claude-attribution.js"

  • PreToolUse

    node "${CLAUDE_PLUGIN_ROOT}/hooks/kill-before-open.js"

View source on GitHub kasempiternal.github.io

Similar plugins

Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).

39 skills

mattpocock/skills 323,108

claude-mem

claudecodex

Persistent memory for Claude Code: the next session starts where the last one left off.

29 skills · 1 MCP server · hooks

thedotmack/claude-mem 116,618

mem0

claudecodex

Cross-session memory and token savings for coding agents.

6 skills · 1 agent · 1 MCP server · hooks

mem0ai/mem0 68,893

context7

claudecodex

Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.

1 MCP server

upstash/context7 64,709

mempalace

claudecodex

AI memory system — mine projects and conversations into a searchable palace. 47 MCP tools, auto-save hooks, guided setup.

3 skills · 6 commands · 1 MCP server · hooks

MemPalace/mempalace 60,536

Automates browser interactions for web testing, form filling, screenshots, and data extraction

12 skills

vercel-labs/agent-browser 50,373

Maintain kstack?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex pluginslisted