AgentationSAFE
The visual feedback tool for agents.
Overview
The visual feedback tool for agents.
471b97d097b2OBSERVED · 2026-10-20Install
Commands as the repository documents them. They are shown, not run.
claude mcp add agentation -- npx -y agentation-mcp server
Host compatibility
What the documentation claims. We have not run a compatibility test.
| Host | Status | Notes |
|---|---|---|
| claude-code | mentioned |
What it tells the agent
The instruction file, verbatim from the audited commit — this is the text the model reads, and the surface the audit's instruction layer examines. Quoted here so you can judge it without cloning anything.
---
name: agentation
description: Add Agentation visual feedback toolbar to a Next.js project
---
# Agentation Setup
Set up the toolbar and, when requested, connect it to the user's coding agent.
A visible toolbar and a healthy MCP server do not prove that browser feedback
reaches the agent. Verify that final step before calling setup complete.
## Steps
1. **Inspect the project**
- Read the project's instructions and detect its package manager.
- Look for `agentation` and an existing `<Agentation>` mount before installing.
- If already mounted, inspect its `endpoint` and current MCP setup. Do not
exit just because the component exists, and do not add a duplicate mount.
- Install `agentation` if missing, using the existing lockfile's package manager.
2. **Choose the connection mode from the request**
- For MCP or live agent feedback, use the same HTTP server URL as the agent.
The default local server is `http://localhost:4747`; preserve an existing
custom port or proxy URL. Never assume the browser and agent use the same
machine when the project is remote.
- For manual copy/paste only, omit `endpoint` and explain that feedback remains
in this browser. Do not silently enable a server connection.
- If the intended mode is unclear, ask one brief question while inspecting
the existing project. Keep working on independent setup checks.
3. **Add or update the component**
For a Next.js App Router project, render in the root layout's body after
children. For Pages Router, render after `Component` in `pages/_app`.
Preserve any existing callbacks and options.
```tsx
import { Agentation } from "agentation";
// Use the actual server URL selected above for MCP sync:
{process.env.NODE_ENV === "development" && (
<Agentation endpoint="http://localhost:4747" />
)}
```
The package provides its own client boundary. Keep the application's root
layout server-rendered unless the project has another reason to change it.
4. **Configure the agent when MCP was requested**
- Use the project's existing MCP client and configuration conventions.
- The launch command must include the `server` subcommand:
`npx -y agentation-mcp server`.
- For supported agents, `npx add-mcp "npx -y agentation-mcp server"` is one
setup option. For Claude Code, use
`claude mcp add agentation -- npx -y agentation-mcp server` or the init wizard.
- Apply the same custom port to the server command and component endpoint.
- Restart or reconnect the agent when its configuration requires it. Do not
create a second HTTP server on a port already owned by the working server.
5. **Verify the complete connection**
- Run `npx agentation-mcp doctor` for service diagnostics. For a custom URL,
use `doctor --http-url <url>` when supported by the installed version.
- Open the app and create one clearly identified test annotation through the
toolbar. Read it with `agentation_get_all_pending` or session-scoped tools.
- Check the exact test comment and page, not just a nonzero count or health
response. Leave unrelated annotations alone.
- If annotations are empty, check the component endpoint, browser network
failures/CORS, and whether the agent and browser point to the same server.
- If browser or MCP tools are unavailable, say exactly which part remains
unverified and give the user this final check. Do not claim live sync works.
## Notes
- The development guard prevents the toolbar from appearing in production.
- Agentation requires React 18 or newer.
- Installing the toolbar does not install, configure or connect an MCP server.
- Manual copy mode is supported and does not need a running server.Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | NA |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (0)
No findings outside the package's declared scope.
Gates applied: no_behavioural_pass.
471b97d097b2full audit observations/trust-audit/skill/benjitaylor__agentation.json · Report an issue / request a re-scanAudit history
Every audit this skill has had.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-20 | 471b97d097b2 | SAFE | B | 89 | first audit |
Questions
What does the agentation — benjitaylor/agentation skill do?
The visual feedback tool for agents.
Is agentation — benjitaylor/agentation safe to install?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean skill reads B.
What can agentation — benjitaylor/agentation access on my machine?
The audit observed no filesystem, network or shell use at all in its source.
What do I need installed to use agentation — benjitaylor/agentation?
Its own instructions reference agentation. Dependencies are pinned to exact versions.
Which assistants does agentation — benjitaylor/agentation work with?
Its documentation mentions claude-code. That is what the text claims, not a compatibility test we ran.
How current is this page?
The grade is for one exact copy of the source (471b97d097b2), read on 2026-10-20. The repository is watched, and a new audit runs when it changes — this is the first audit.
Add this badge to your README
Shows this page's audit result and links back to it. It updates when the audit does.
[](https://modelsatlas.com/skills/benjitaylor/agentation/)
Not your listing, or something out of date? Claim or update your listing.