HUMANMIRROR PRO

Security history and resource control for AI-built systems.

Run the scanner locally. Upload only findings metadata. Track repositories, API quotas, token budgets and operational needs from one place.

npx --yes https://humanmirror.fr/downloads/humanmirror-audit-0.2.0.tgz

Open your workspace

Paste a HumanMirror Pro API key, or start the €49/month plan by card.

API key Stripe Checkout · card payment · cancel anytime from Stripe billing support.

PRO WORKSPACE

Engineering control plane

ACTIVE

Save this API key now.It is shown once. Store it in your secret manager.

Latest security score—No cloud audit yet

Repositories—Cloud history

API remaining——

Needs attention—Resource thresholds

VIBE-CODE SECURITY

Audit history

The CLI scans source locally. Cloud history stores scores, rule IDs, paths and line numbers — not source contents.

Repo	Score	Risk	Findings	When

USAGE

HumanMirror API

——

X-API-Key: hm_live_…
POST /api/v1/m2m/resource/

FLOW GUARD

Prompt-injection filter for n8n, Make & webhooks

Place this HTTP call immediately before an LLM or tool-execution node. Flow Guard returns a verdict, risk score, sanitized payload and signed cleanliness proof.

POST /api/v1/saas/guard/
X-API-Key: hm_live_…
Content-Type: application/json

{"text":"untrusted input from email/form/webhook"}

PRO UTILITY

One subscription, two operational controls.

HumanMirror Pro includes the Vibe-Code security history, Flow Guard filtering, team keys, Stocks & Needs and 100,000 API calls/month.

€49 / month · Stripe card
No wallet required
Enterprise x402 remains separate

M2M NEXT-GEN · PREVIEW 2027–2030

Pack Futur, included in HumanMirror Pro Preview.

Test the next control layer for autonomous-agent fleets with the same Pro API key. Calls use your existing monthly API quota; Enterprise Fleet includes unlimited Preview access.

INTENT PROOF

Passport of Intent

Analyze untrusted payloads, detect prompt-injection risk and issue a signed safe-routing certificate. The non-malice score is a risk signal, not a guarantee.

POST /v1/agent/intent-proof/
X-API-Key: hm_live_…
{"payload":{"instruction":"untrusted agent input"}}

FLEET TRIBUNAL

Consensus Lock

Arbitrate 2–12 model decisions with deterministic confidence/weight scoring and receive a signed execution receipt.

POST /v1/fleet/consensus-lock/
X-API-Key: hm_live_…
{"models":[{"id":"a","decision":"APPROVE","confidence":0.8},{"id":"b","decision":"APPROVE","confidence":0.7}]}

SMART ESCROW

Escrow Settle

Evaluate delivery-quality checks and issue a signed RELEASE, REFUND or REVIEW decision. Preview is non-custodial: HumanMirror does not hold or move third-party funds.

POST /v1/m2m/escrow-settle/
X-API-Key: hm_live_…
{"escrow_reference":"order-123","checks":[{"name":"quality","score":0.92}]}

STOCKS & NEEDS

Know what is running out before production does.

Track API credits, model tokens, cloud budgets or any technical resource. When a threshold is crossed, HumanMirror records the need and can call an HTTPS webhook.

Stocks

Needs & alerts

TEAM

Workspace keys

Create separate API keys for teammates or CI. New secrets are shown once and can be revoked independently.

DATA BOUNDARY

Local-first by design.

Source code remains on your machine unless you explicitly send it elsewhere.
Audit cloud history stores finding metadata, hashes/identifiers and scores.
API keys are hashed server-side; one-time claim material is encrypted until claimed.
Resource webhooks reject local/private network destinations.