Gildara MCP server

by gildara.io·io.gildara/mcp-server·v0.8.0

No reviews yet

Operating contracts for AI agents. Structured instructions compiled into system prompts.

B82/100grade B

Write a review Install

What users say

No reviews yet

Be the first

Safety scan

B82/100

full report

Adoption

New

0 stars45 downloads/wk

Reviews

Write one

Nobody has reviewed Gildara yet.

If you have run it, two minutes of your experience saves the next person an afternoon.

Gildara tools (12, 3 write)

write = sends, deletes, buys or posts

Read from the package source without running it. The installed server may list more.

append_memory
Append content to an existing prompt or memory. Creates a new version with the appended content added to the end. Use this for accumulating memories, adding context, or building up a prompt incrementally. The category 'memory' is recommended for user context that should be portable across AI tools.
create_promptwrite action
Create a new prompt in your Gildara vault. Returns the new prompt ID which you can then use with resolve_prompt or run_prompt.
get_account_link
Get the URL to link this agent's vault to a human Gildara account (unlocks sync across devices and the web UI at gildara.io). Returns the pairing code and link URL — show both to the user so they can open the URL in a browser. Use this when the user asks how to log in, pair, claim, or link this agent, or right after first-run auto-provisioning.
get_brief_result
Fetch the result an agent submitted for a previously-dispatched Brief.
get_prompt
Get details of a specific prompt including its content, variables, tags, and operating contract configuration.
list_blueprints
List all available agent blueprint templates. These are pre-built operating contracts for common agent types (code review, legal analysis, triage, etc.) that you can browse and use as starting points. Supports search by keyword or category.
list_prompts
List all prompts in your Gildara vault. Returns titles, IDs, categories, and whether each has an operating contract enabled. Use search_prompts instead when you want to find prompts by topic or concept rather than listing everything.
list_returned_briefs
List briefs that have results returned from their agents (status='returned'),
resolve_prompt
Resolve a prompt into its compiled system prompt, with operating contract sections assembled (role, tools, stop conditions, output schema). This is what you should pass as the system prompt to an AI model. Supports variable substitution and channel selection (latest/stable).
run_promptwrite action
Run a prompt through AI with automatic output validation and auto-repair. If the prompt has an output contract (JSON schema), the response is validated and automatically retried with a repair prompt on failure. Returns the AI response, parsed JSON (if applicable), and validation metadata.
save_briefwrite action
Create a Brief in Gildara and dispatch it to a named agent.
search_prompts
Search your Gildara vault semantically by meaning, not just keywords. Returns the most relevant prompts for a natural-language query (e.g. 'reviewing code for security issues', 'summarizing customer feedback'). Ranked by vector similarity over title, description, category, tags, and content. Use this first before list_prompts when looking for a prompt that matches a concept.

Public scan report

scanner v0.1.9 · 2026-09-24 · same rubric, same numbers if you re-run it

no findings

Code scan9 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 qualitylocal package, no credentials required12/15
Maintenancelast push 60 days ago12/15
Maintainer identitynamespace and repository owner differ4/10

Overall 82/100. Components that don't apply are left out of the denominator. Any critical finding is an F.Rubric Appeal a finding JSON

How Gildara compares

Against every graded MCP server, from our own scans and uptime checks. Updated after each scan run.

Safety score: 82/100; higher than 63% of 48,255 servers
Tools: 12; the typical server has 7
Tools that change things: 25%; typical: 0%. Fewer write tools means less an agent can break
Downloads a week: 45; more than 37% of 12,191 servers with a package

What its tools let an agent do: 9 read · 2 create · 1 run or deploy

Install Gildara in Claude Code, Cursor or VS Code

Runs npx -y @gildara/mcp-server on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add mcp-server -- npx -y @gildara/mcp-server

Add to Cursor

What the publisher says

From the Gildara repository's README, as published. We do not edit it. Read it on GitHub

Gildara MCP

One prompt vault. Every agent. Zero setup. Give Claude, Cursor, Windsurf, and Claude Code a shared, searchable prompt library that survives across chats, machines, and tools.

{ "mcpServers": { "gildara": { "command": "npx", "args": ["-y", "@gildara/mcp-server"] } } }

Paste into claudedesktopconfig.json, restart Claude — your Gildara tools appear in Claude's tool list. No signup. No API key. The server auto-provisions a free agent on first run and prints a one-click link code to pair it with your account later.

→ Get started at gildara.io — pair your auto-provisioned agent to sync your vault and unlock Pro.

What your AI can do

Search prompts semantically — find by meaning, not just keywords (Gemini embeddings).
Resolve + run prompts — compiled system prompts with operating-contract sections assembled, auto-repair on JSON schema failures.
Create + append — save new prompts, or append persistent facts to memory-type prompts.
Dispatch briefs to agents — say "send this to gevorg" and the MCP server creates a Brief, ships it as a .md attachment to your paired Telegram chat, and @mentions the target agent's bot. One call closes the sketch-in-chat → agent-on-another-box → result-in-chat loop. See docs/briefs/dogfood-rollout-v0.md for the agent-side integration recipe.
Browse 48+ blueprints — pre-built agent templates.

Setup

No API key required on first run. As of v0.6.0, the MCP server auto-provisions a new agent account on first startup if no GILDARAAPIKEY is set, and caches the key at ~/.gildara/auto-key.json for future runs. Skip to "Add to Claude Desktop" below — zero curl required.

On first run, the server prints a link code to stderr (visible in Claude Desktop's MCP logs). Visit https://gildara.io/link?code=XXXXXX to pair the auto-provisioned agent with your human account and inherit your subscription tier.

1. (Optional) Use an existing API key instead of auto-provisioning

Skip this whole section if you're happy with auto-provisioning.

Otherwise, get a key one of these ways:

Manual provision via API:

curl -X POST https://gildara.io/api/v1/provision \
  -H 'Content-Type: application/json' \
  -d '{"agent_label": "my-agent"}'

From the web: Go to gildara.io/account and create an API key with the agent-standard preset.

Then set GILDARAAPIKEY in your MCP config (see examples below).

2. Add to Claude Desktop

Edit your claudedesktopconfig.json. No env block needed for auto-provision:

{
  "mcpServers": {
    "gildara": {
      "command": "npx",
      "args": ["-y", "@gildara/mcp-server"]
    }
  }
}

Or, with an existing key:

{
  "mcpServers": {
    "gildara": {
      "command": "npx",
      "args": ["-y", "@gildara/mcp-server"],
      "env": {
        "GILDARA_API_KEY": "pvk_your_key_here"
      }
    }
  }
}

3. Add to Cursor

One click:

Or manually: edit .cursor/mcp.json in your project root using the same configuration as Claude Desktop (with or without the env block — same rules apply).

4. Add to Claude Code

Run this in your terminal:

claude mcp add gildara -- npx -y @gildara/mcp-server

Shortened. The full README is on GitHub.

Nothing above is checked by us. What we check is on the safety report.

Gildara: common questions

Is Gildara MCP server safe?: Mostly: it is graded B (82/100). Read the Gildara safety report
How do I install Gildara?: It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does Gildara need an API key?: No secret keys are declared. It reads 1 setting from the environment.
Is Gildara maintained?: The last commit was 100 days ago (2026-07-27). The latest release is v0.8.0.
How does Gildara compare with other MCP servers?: safety score 82/100 (higher than 63% of 48,255 servers); tools 12 (the typical server has 7); tools that change things 25% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.