Ctrl/
Provides trust assessments, cryptographic verification, and risk analysis for AI systems.
Input:
Output:
Updated: Feb 24, 2026 Free + from $0.01/month
Description
Mnemom offers a verification layer for AI agents, encompassing identity, integrity analysis, cryptographic validation, reputation scoring, risk assessment, and containment in a unified stack. By connecting your agent to the gateway (a single environment variable, requiring no code alterations), every action is tracked, analyzed, and cryptographically verified using Ed25519 signatures, SHA-256 hash chains, Merkle trees, and zero-knowledge proofs.
Each agent receives a Trust Score, a rating from AAA to CCC, calculated using five weighted indicators. Fleet operators benefit from coherence analysis, team risk modeling (CoVaR, Markowitz, DebtRank), and real-time containment with automated pause/kill/resume functionalities.
Compatible with Anthropic, OpenAI, and Gemini. Utilizes open protocols (AAP/AIP) on npm and PyPI. Licensed under Apache.
Each agent receives a Trust Score, a rating from AAA to CCC, calculated using five weighted indicators. Fleet operators benefit from coherence analysis, team risk modeling (CoVaR, Markowitz, DebtRank), and real-time containment with automated pause/kill/resume functionalities.
Compatible with Anthropic, OpenAI, and Gemini. Utilizes open protocols (AAP/AIP) on npm and PyPI. Licensed under Apache.
Key Features
Agent Reputation Bond Rating (aaa-ccc)
Cryptographic Integrity Certificates (ed25519, Hash Chained, Merkle Trees, Zk Proofs)
Real-time Agent Containment Engine (auto-pause kill resume)
Team Risk Assessment (covar, Markowitz, Debtrank)
Drop-in Gateway — Zero Agent Code Changes
Pricing Plans
Model
freemium
Packages
1 Package
Price Start From
$0.01/month
Payment Model
monthly
Model
freemium
Packages
1 Package
Price Start From
$0.01/month
Payment Model
monthly
Releases
- Trust Score™: Ratings (AAA–CCC) for each AI agent, derived from five weighted signals: integrity, compliance, drift, trace quality, and coherence
- Zero-knowledge proofs: Every agent action gains a cryptographic attestation chain — Ed25519 signatures, SHA-256 hash chains, Merkle trees, SP1 STARK proofs
- Team risk engine: CoVaR, Markowitz portfolio theory, and DebtRank applied to multi-agent fleets — identify which agent is negatively impacting quality
- Real-time containment: Automatically pause, stop, or restart agents when trust falls below a specified level
Smoltbot gateway: One environment variable, zero code changes — supports Anthropic, OpenAI, and Gemini
- Open protocols: AAP (behavioral contracts) and AIP (runtime analysis) on npm and PyPI
- Zero-knowledge proofs: Every agent action gains a cryptographic attestation chain — Ed25519 signatures, SHA-256 hash chains, Merkle trees, SP1 STARK proofs
- Team risk engine: CoVaR, Markowitz portfolio theory, and DebtRank applied to multi-agent fleets — identify which agent is negatively impacting quality
- Real-time containment: Automatically pause, stop, or restart agents when trust falls below a specified level
Smoltbot gateway: One environment variable, zero code changes — supports Anthropic, OpenAI, and Gemini
- Open protocols: AAP (behavioral contracts) and AIP (runtime analysis) on npm and PyPI
Reviews
Pros & Cons
Pros
Agent reputation bond rating (AAA-CCC)
Functions with Anthropic, OpenAI, Gemini
Real-time containment (auto pause kill resume)
Cons
Requires API key per provider
No native mobile app
No offline operability
