Skip to content
verify mcp Beta VerifyMCP is currently in beta. If you notice any issues, get in touch and we’ll put it right.

Marigold

NPM · MARIGOLD-DRAFT · 2 COMPONENTS · SCANNED SEP 20

Comment on AI-generated webpages; feedback flows back to your coding agent. Free, MIT, local-first.

−1 this week 80 Trust /100
Trust breakdown (7 categories)

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. How we score → Why this is hard to score →

Supply Chain Security100
  • No malware found by supply-chain analysis.Pass
  • No known CVEs affecting this package version or its production dependencies.Pass
  • No install/post-install scripts declared.Pass
  • No production dependencies, so there is no dependency health to assess. View diagnostics → Pass
Provenance & Transparency45
Schema Quality & AI Usability74
  • AI-judged instruction clarity (excellent).Pass
  • Context-footprint check failed: tool/resource definitions use about 1705 tokens (~131/item across 13 items; 13 tools + 0 resources), over budget; trim descriptions and params. See how to fix → Fail
  • Usage-examples check failed: none of the tools include examples. See how to fix → Fail
Stability & Change Management80
  • Stability observed for 24 of 30 days with no destabilising changes; credit accrues until the full window elapses.Partial
Tool Coverage82
  • 100% of tools have a non-trivial description (not blank, and not just the tool's name).Pass
  • 47% of tool parameters carry a description.Partial
Tool Safety100
  • No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.Pass
  • We read all 13 captured tool definition(s), and no name or description among them implies an irreversible operation.Pass
  • An AI judge read all 14 captured unit(s) of tool text and found none that tries to manipulate the model reading it.Pass
Capabilities100
  • Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.Pass
Install

How do I install the Marigold MCP server?

Marigold runs locally as an npm package, launched with npx -y marigold-draft. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

npm · marigold-draft

# add to Claude Code
claude mcp add immuneeb-marigold -- npx -y marigold-draft
// .cursor/mcp.json
{
  "mcpServers": {
    "immuneeb-marigold": {
      "command": "npx",
      "args": [
        "-y",
        "marigold-draft"
      ]
    }
  }
}
// .vscode/mcp.json
{
  "servers": {
    "immuneeb-marigold": {
      "command": "npx",
      "args": [
        "-y",
        "marigold-draft"
      ]
    }
  }
}
# add to Codex CLI
codex mcp add immuneeb-marigold -- npx -y marigold-draft
// opencode.json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "immuneeb-marigold": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "marigold-draft"
      ],
      "enabled": true
    }
  }
}
# add to OpenClaw
openclaw mcp add immuneeb-marigold --command npx --arg -y --arg marigold-draft
# ~/.hermes/config.yaml
mcp_servers:
  immuneeb-marigold:
    command: "npx"
    args: ["-y", "marigold-draft"]
// ~/.netclaw/config/netclaw.json
{
  "McpServers": {
    "immuneeb-marigold": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "marigold-draft"
      ]
    }
  }
}
# add to Vellum
assistant mcp add immuneeb-marigold -t stdio -c npx -a -y marigold-draft
// mcp.json
{
  "mcpServers": {
    "immuneeb-marigold": {
      "command": "npx",
      "args": [
        "-y",
        "marigold-draft"
      ]
    }
  }
}
Changelog

Every change we have recorded for this component, newest first. Security-relevant changes are always shown. ▲ marks a change for the better, ▼ a change for the worse; unmarked changes are neutral.

  • 19 Sept 26 −3
    • Stability: pass → 0.77 functional
  • 18 Sept 26 +1
    • Stability: 0.97 → pass security
  • 16 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 90 to 93. That category is still filling its 30-day observation window: 27 days of observed history at the previous scan, 28 at this one. The score rises as the window fills, whether or not the server changes.

  • 13 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 80 to 83. That category is still filling its 30-day observation window: 24 days of observed history at the previous scan, 25 at this one. The score rises as the window fills, whether or not the server changes.

  • 12 Sept 26 −3
    • Stability: pass → 0.80 functional
  • 11 Sept 26 +1
    • Stability: 0.97 → pass security
  • 9 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 90 to 93. That category is still filling its 30-day observation window: 27 days of observed history at the previous scan, 28 at this one. The score rises as the window fills, whether or not the server changes.

  • 6 Sept 26 +1

    No change was recorded against any check on this day. Stability & Change Management went from 80 to 83. That category is still filling its 30-day observation window: 24 days of observed history at the previous scan, 25 at this one. The score rises as the window fills, whether or not the server changes.

Diagnostics

Diagnostic detail from the automated scan of this channel: what the scanner observed at each step, so you can see exactly where a check passed or failed. It is informational only and never changes the trust score.

Captured 20 Sept 2026 · Analysed npm/marigold-draft@0.9.1

Provenance No attestation

The registry publishes no build provenance for this version, so there is nothing to verify.

Result No attestation
Ecosystem npm

Background: How many MCP packages publish verified provenance →

Dependencies 0 packages
Packages resolved 0
Tree resolution Complete

Background: SBOMs and build attestations, explained →

MCP tools · 13 exposed · ~1,311 tokens

The tools this component advertises to a client, with an estimated token cost for each. Expand a tool to see its parameters and schema. The per-tool counts are indicative and are not scored directly; the schema's total context footprint is one signal in Schema Quality & AI Usability. A tool's description is untrusted text the model reads on every call, which is what makes this list a security surface and not just an inventory: how tool poisoning works →

Tool Tokens
create_draft ~92

write a new HTML/SVG draft to disk and open it in the reviewer's browser. Returns the file path — pass it to every other tool. Use update_draft for revisions.

NameTypeReqDescription
htmlstringyesFull HTML document, fragment, or SVG markup
namestringFile name without extension (defaults to the title)
titlestringDoc title (also names the file)

No output schema declared.

No examples provided.

get_context ~176

get the draft's catch-up digest: durable insights FIRST (owner-level learnings across all drafts, stale ones first — an insight goes stale when a cited thread gets fresh REVIEWER activity; reaffirm/refine before creating a near-duplicate), then the open comments, the recent changes, and this draft's unsynthesized episodes (each thread's full comment chain, every attempt, and where it landed — read the whole chain, a reopen may be a follow-up, not a rejection). An episode's terminalState is confirmed | proposed | open | "answered" (an open Q&A thread you replied to that was never resolved or reopened — synthesizable like any other). Answered threads appear BOTH as episodes and as open comments carrying `answered:true` — dedupe the two views by that flag.

NameTypeReqDescription
pathstringyes

No output schema declared.

No examples provided.

get_feedback ~88

read the draft's comments. With waitSeconds > 0, wait for the reviewer to send feedback (or time out) and return the review round — use this after creating or updating a draft to wait for the reviewer.

NameTypeReqDescription
pathstringyes
waitSecondsnumberHow long to block for the next review round; 0/omitted = return current comments now

No output schema declared.

No examples provided.

get_history ~70

list the draft's recent changes, most recent first — each with its version, actor, intent, diff stats, and a few changed/added/removed element summaries.

NameTypeReqDescription
limitnumberHow many recent changes to return (default 50, max 200)
pathstringyes

No output schema declared.

No examples provided.

get_insight ~46

expand one insight by id: its statement, status, evidence links, and the full review episode behind each cited comment (chain + attempts + outcome).

NameTypeReqDescription
idstringyes

No output schema declared.

No examples provided.

note_intent ~70

record the intent — the "why" — for the draft's next save. The next edit's change entry carries it, so the history reads as decisions, not just diffs.

NameTypeReqDescription
intentstringyesOne line: why the next edit is being made
pathstringyes

No output schema declared.

No examples provided.

open_draft ~49

open an existing .html/.htm/.svg draft in the review shell (starts the daemon if needed).

NameTypeReqDescription
pathstringyesPath to the file
titlestring

No output schema declared.

No examples provided.

read_draft ~36

read the draft's current HTML source from disk (e.g. before an update_draft revision).

NameTypeReqDescription
pathstringyes

No output schema declared.

No examples provided.

reply_to_comment ~50

reply in a comment thread — shown with the AI badge. Say what you changed before resolving.

NameTypeReqDescription
bodystringyes
commentIdstringyes
pathstringyes

No output schema declared.

No examples provided.

resolve_comment ~81

propose a comment thread resolved, or reopen it. Resolving is a PROPOSAL — the reviewer confirms it (or reopens it if the fix missed) in the shell; it is not final until they do. Reply first with what you changed.

NameTypeReqDescription
commentIdstringyes
pathstringyes
reopenboolean

No output schema declared.

No examples provided.

save_insight ~351

record a durable, cross-draft learning distilled from review episodes — owner-level, not per-draft. Evidence is mandatory: cite the {docId, commentId} threads it came from. Statements are capped at 140 characters. relation: "new" creates; "reinforces"/"refines"/"contradicts" update the insight named by `updates`. Creating (or refining into a) statement close to an existing ACTIVE insight is refused with `{saved:false, needsDistinction:true, candidates:[{id,statement}], hint}` — reaffirm one (updates + relation:"reinforces"), refine it (relation:"refines" + a new statement, which supersedes the old), contradict it (relation:"contradicts" — flags it, keeps it), or resend with distinctFrom listing those candidate ids. On success: `{saved:true, insight:{id,statement,status,relation,evidenceCount,updatedAt}, supersededId?}`. Distill at a review round's close.

NameTypeReqDescription
createdByLabelstring
distinctFromarrayids of similar active insights you've decided are genuinely distinct — clears the forced choice (new + refines)
evidencearrayyes≥1 {docId, commentId} pairs that name real comments in a draft
relationstringnew = create; reinforces = add evidence to `updates`; refines = new insight supersedes `updates`; contradicts = flag `updates` false
statementstringThe learning, one sentence (≤140 chars). Required for new + refines.
updatesstringThe existing insight this reinforces/refines/contradicts

No output schema declared.

No examples provided.

start_analysis ~147

load the Marigold Way before authoring a draft — the first-principles method, doc structure guide, and a mode posture pack for what the session must produce. Call this first when asked to analyze, explain, teach, or build an interactive draft, then follow the returned method.

NameTypeReqDescription
modestringWhat the session must produce: analyze = first-principles breakdown (default); learn = a retained mental model; judge = verdicts on existing work; decide = a selection + rationale; organize = an arra…
topicstringThe topic, if known

No output schema declared.

No examples provided.

update_draft ~55

replace the draft's content with revised HTML. The reviewer's open tab live-reloads and existing comments re-anchor (keep the DOM structure stable where you can).

NameTypeReqDescription
htmlstringyes
pathstringyes

No output schema declared.

No examples provided.

Common questions

What is the Marigold MCP server?

Marigold is an MCP server listed in the public MCP registry as io.github.immuneeb/marigold. Comment on AI-generated webpages; feedback flows back to your coding agent. Free, MIT, local-first. This page covers its npm package (marigold-draft).

Is the Marigold MCP server safe to use?

Marigold scores 80 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 20 September 2026. It declares no install or post-install scripts. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

What tools does the Marigold MCP server expose?

Marigold exposes 13 tools: start_analysis, create_draft, open_draft, update_draft, get_feedback, and 8 more. Their descriptions and schemas cost roughly 1,311 tokens of context every time the server is loaded.

Is the Marigold MCP server still maintained?

Marigold is still listed as active in the MCP registry. We last reached this channel on 20 September 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

What licence is the Marigold MCP server under?

Marigold declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.