Guardian
Deterministic recipe verification for AI agents. Catches hallucinated temperatures, missing techniques, wrong ingredients, and impossible cooking steps — then suggests one-click fixes. Covers 161+ dishes across 5 cuisines. Also checks EU allergen compliance (all 14 Annex II groups), verifies dietary claims (vegan, halal, kosher, gluten-free), and flags safety issues. Verify against Guardian's built-in master specs or your own house recipe. Same input → same verdict, every time. No API key required.
How to connect
Same URL, five paste targets. Ready surface · Auth required. Not a safe-to-install badge. Handshake this URL
Paste into Settings → MCP or .cursor/mcp.json.
{
"mcpServers": {
"kaimeilabs-guardian-engine": {
"url": "https://guardian-engine--kaimeilabs.run.tools"
}
}
}
Settings → Connectors → Add → Custom → Web. Paste the URL. Enable in a new chat.
Claude → Settings → Connectors → Add → Custom → Web
Name: Guardian
URL: https://guardian-engine--kaimeilabs.run.tools
OAuth only if this server requires it.
Save, start a new chat, enable the connector.
Developer mode custom connector. ChatGPT cannot reach localhost.
Settings → Apps & Connectors → Advanced → Developer mode ON
Create custom connector
Name: Guardian
MCP server URL: https://guardian-engine--kaimeilabs.run.tools
Auth: none unless the server requires OAuth
New chat → + → Developer mode → enable the connector
Register the HTTPS URL, then claude mcp list.
claude mcp add --transport http kaimeilabs-guardian-engine \
https://guardian-engine--kaimeilabs.run.tools
Pass on session/new as mcpServers when the agent advertised HTTP MCP.
[
{
"type": "http",
"name": "kaimeilabs-guardian-engine",
"url": "https://guardian-engine--kaimeilabs.run.tools"
}
]
Paste into a README or docs site. The widget loads snippets from Influzer — still not a SAFE badge.
Same handshake clock as this listing (overlay / warm / seed; live tools/list only if that is missing). Indigo is live_ok, amber is auth_required, gray is unreachable. Not a SAFE badge. Stdio listings have no badge.
[](https://www.influzer.ai/mcp/probe?url=https%3A%2F%2Fguardian-engine--kaimeilabs.run.tools)
<div data-influzer-install="kaimeilabs-guardian-engine"></div>
<script async src="https://www.influzer.ai/embed/install.js"></script>
https://www.influzer.ai/embed/mcp/kaimeilabs-guardian-engine.md
JSON: https://www.influzer.ai/api/v1/install/kaimeilabs-guardian-engine
Recipes using this server
No published recipes yet for this server.
Listing quality
Observable facts from our catalog and daily probe — not a safe-to-install badge.
- Surface
- Ready surface
- Live handshake
- Auth required
- Auth gate
- Auth gate
- Tools source
- tools/list (live)
Observable catalog + probe facts — not a safe-to-install badge.
Tools
(7)-
verify_recipeVerify a candidate recipe against a Guardian master recipe. Uses deterministic graph-based verification to check technique, temperature, timing, cooking medium, and required ingredients. **Verdict**: `verdict` is strictly PASSED or FAILED and is policy-driven — any CRITICAL finding fails the recipe; more than 5 WARNINGs also fail. There is no score in the response (ADR-013): gate on `verdict` and explain failures from `findings`. **Field audience**: `issue` is a machine-readable code for prog
-
fix_recipeDeterministically repair a candidate recipe against a Guardian master. Verifies the candidate, applies every machine-actionable correction the symbolic engine produced (missing ingredients, quantities, temperatures, durations, cooking media, ingredient substitutions), then re-verifies the result. No LLM is used — the repair is a deterministic function of the candidate recipe and the master ruleset. Findings that need recipe-authoring judgement — adding a whole cooking phase, rewriting step ins
-
list_dishesList all available master dishes with rich metadata. Returns: Dictionary with `schema_version` and a `dishes` list. Each dish includes slug, title, cuisine, region, aliases, and complexity.
-
get_masterReturn the canonical master recipe for a dish (read-only, no LLM). Enables compare-then-verify agentic loops: fetch the master, diff it against the user's recipe, then call verify_recipe — instead of verifying blind. Pure knowledge-base lookup, no LLM in the hot path. Master content is transparent by default (ADR-009 / ADR-010): exact temperatures, timings, and EU FIC 1169/2011 allergen codes are returned verbatim, never obfuscated. No score is included (ADR-013) — this is reference data, not
-
check_safetyRun master-independent safety checks on a candidate recipe. Works for ANY recipe — no dish resolution, no master SOP required. Checks poultry internal-temperature safety and scans all ingredients for the 14 EU FIC 1169/2011 Annex II allergen groups. The verdict is a deterministic function of (candidate, kb_version_hash) — no LLM involvement. Use this when verify_recipe has no matching master for the dish: the safety layer still applies to every recipe. Returns: Safety envelope: verdict (P
-
check_allergensCheck ingredients for EU FIC 1169/2011 allergen compliance. Returns a detailed audit trace mapping each ingredient to its EU Annex II allergen group with entry numbers and labels. The safety verdict is deterministic — no LLM involvement in the decision. Use check_all_eu_allergens=True for food labelling (detect all allergens). Use restrictions=['dairy', 'gluten'] to check for specific user allergies.
-
verify_dietary_claimVerify that a recipe satisfies a dietary claim (vegan, halal, gluten-free, ...). Reuses the existing allergen-detection logic plus a curated forbidden-ingredient map (apps/guardian/knowledge/dietary_claims.yaml). Returns a structured verdict with the specific offending ingredients and a short justification — never a vague paraphrase.
Tool source: tools/list (live) · last probed 2026-10-24