Tyrekick MCP server
Read and resolve human feedback pinned on your live prototype, straight from your coding agent.
3 stars32 downloads/wk
Reviews
Write oneNobody has reviewed Tyrekick yet.
If you have run it, two minutes of your experience saves the next person an afternoon.
Tyrekick tools (6)
write = sends, deletes, buys or postsRead from the package source without running it. The installed server may list more.
feedback_statsAggregate counts of feedback by status, route, and app_version
get_feedbackFetch the full feedback record by id, including the anchor
list_feedbackList reviewer feedback collected by the Tyrekick worker, newest first.
resolve_feedbackMark a feedback item as resolved (optionally with a note describing the fix).
retrospectiveThe AI feedback loop: analyse your own feedback history for patterns — what
triage_feedbackMove an open feedback item through triage: approve it (cleared for an
Public scan report
scanner v0.1.9 · 2026-09-23 · same rubric, same numbers if you re-run it
- Code scan6 source files scanned25/25
- –Live reliabilityno gateway calls yet and no remote to proben/a
- –Tool poisoningtools not inspected (local package is not executed); not countedn/a
- Auth qualitystatic API keys via environment variables6/15
- Maintenancelast push 7 days ago15/15
- Maintainer identityregistry namespace matches repository owner; GitHub account older than a year8/10
What the publisher says
From the Tyrekick repository's README, as published. We do not edit it. Read it on GitHub
Tyrekick
Kick the tyres before you ship it. Get people you trust to review your AI-built prototype — and let your coding agent act on what they say.
Reviewers pin comments directly on the live page — no account, no training. Your coding agent pulls them back over MCP with the exact element, its visible text, the section heading, and any page errors attached, fixes what was flagged, and marks it resolved. The reviewer's pin turns green with a note. You never retype feedback into a prompt again.
▶ Try the live demo · Quickstart · Docs · "The Outer Loop" — why this exists
Real footage: that comment was pinned on the live demo, fixed by an agent, and the pin turned green — no retyping, no tracker, no accounts.
How it works
reviewer's browser destination YOU own your coding agent
┌──────────────────┐ POST ┌────────────────────┐ MCP ┌─────────────────┐
│ pin a comment on │ ────────► │ Discord webhook or │ ◄───────► │ "fix the open │
│ the live page │ │ your CF Worker │ │ feedback" │
└──────────────────┘ └────────────────────┘ └─────────────────┘The widget is a single ~14 KB (gzipped) script with zero dependencies, rendered in Shadow DOM so it can't fight your page. There is no Tyrekick backend, no accounts, and nothing phones home — feedback POSTs straight from the reviewer's browser to a destination you own. It exists for the moment an agent-built prototype needs more than one pair of eyes: instead of pasting screenshots into group chats and losing the replies, every comment arrives structured, versioned, and pinned to the exact spot.
Pins stay on the page after you leave a comment (dimmed until you interact) — hover one to see its text, click it to open the thread right there, with reply, retry, and discard actions at the pin. Reviewers also get a comment drawer (a right-hand overview listing every pin with its text — click one to jump back to that spot, or hide pins entirely), draft recovery, and full keyboard/touch support (Cmd/Ctrl+Enter sends). With persist: true (the default), localStorage is only used to recover unsent work across reloads — comments that couldn't be delivered come back after a reload with Retry and Discard buttons, re-pinned to the element they were about even if the window size changed.
Agents: installing Tyrekick into a project? Use the make-reviewable skill — one sentence and it hosts the page if needed, picks a destination, deploys the worker, installs the widget, wires MCP, and drafts the ask to send reviewers — or follow AGENTS.md step by step.
Full documentation lives in docs/: quickstart · getting started · configuration · the reviewer experience · destinations · the agent loop · payload reference · troubleshooting & FAQ
Sharing a prototype publicly? Start with taking a prototype public, then the per-capability guides: rate limiting · shared review · AI auto-reply.
Shortened. The full README is on GitHub.
Nothing above is checked by us. What we check is on the safety report.
Install directly
Runs npx -y tyrekick-mcp on your machine. Read the scan report first; the gateway never runs local packages.
claude mcp add tyrekick -- npx -y tyrekick-mcp
Tyrekick: common questions
- Is Tyrekick MCP server safe?
- Mostly: it is graded B (83/100). Read the Tyrekick safety report
- How do I install Tyrekick?
- It runs on your machine. Copy the Claude Code, Claude Desktop or Cursor config from the install section.
- Does Tyrekick need an API key?
- Yes. The registry entry asks for
TYREKICK_TOKEN. - Is Tyrekick maintained?
- The last commit was 9 days ago (2026-09-17). The latest release is v0.3.1.
- What can I use instead of Tyrekick?
- Servers from other publishers that do the same job: Doc Bridge MCP server, Agent-Native Plan MCP server and Godot MCP server. Compare all Tyrekick alternatives.
Alternatives to Tyrekick
Same job from other publishers: the closest match first, then the best rated.
- Doc BridgeResolve documentation ownership, edit roots, checks, and human handoffs before an agent edits code.not reviewedGrowingA
- Agent-Native PlanStructured visual plans and PR recaps with diagrams, prototypes, annotations, and sharingnot reviewedEstablishedA
- Godot MCPAgent-driven Godot playtesting: editor control, input injection, game-time stepping, live state.not reviewedEstablishedA
- OrcareplayRead, replay and fork recorded coding-agent runs.not reviewedEstablishedA
- three.ws Avatarsthree.ws 3D avatars in your agent: render a live, interactive 3D avatar inline, or get an embed.not reviewedEstablishedA