MCP.Market

SWT3 AI Witness MCP server

by tenova.io·io.tenova/swt3-witness·v0.7.5
No reviews yet

Cryptographic AI governance and audit. 75 tools, 84 frameworks. EU AI Act, NIST, CMMC, MCP Top 10.

B80/100grade B
Write a reviewInstall
What users say
No reviews yet
Be the first
Safety scan
B80/100

full report

Adoption
Growing

0 stars670 downloads/wk

Reviews

Write one

Nobody has reviewed SWT3 AI Witness yet.

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

SWT3 AI Witness tools (75, 1 write)

write = sends, deletes, buys or posts

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

  • attest_memory_context

    Attest persistent memory sources influencing decisions (AI-SKILL.2).

  • attest_skill_manifest

    Attest the active skill/tool/plugin manifest (AI-SKILL.1).

  • chain_handoff

    Witness a handoff between agents in a multi-agent chain.

  • check_posture

    Check the current AI witness compliance posture for your tenant.

  • coverage_report

    Report framework coverage for the current audit session. Shows which

  • end_audit_session

    End the audit session and produce a compliance gap report.

  • gate_evaluate

    Parse and validate a .swt3-gate.yml governance gate configuration.

  • list_procedures

    List available UCT (Universal Control Taxonomy) procedures from the SWT3 registry.

  • present_trust_credential

    Get this agent's trust credential for presentation to another agent.

  • reconstruct_timeline

    Reconstruct a forensic timeline of SWT3 witness anchors.

  • report_violation

    Voluntarily self-report a policy violation.

  • resolve_crosswalk

    Look up regulatory crosswalk mappings. Given a procedure ID (e.g., AI-FAIR.1),

  • resolve_jurisdiction

    Look up which regulatory frameworks apply to a jurisdiction. Given an ISO 3166-1 country code

  • signup

    Get a signup link to create a free SWT3 account.

  • start_audit_session

    Begin passive compliance tracking for this conversation.

  • start_chain

    Generate a cycle_id for a multi-agent chain.

  • suggest_procedures

    Get advisory suggestions for which SWT3 procedures to witness based on context.

  • verify_agent_trust

    Verify a counterpart agent's compliance posture before exchanging data or calling their tools (AI-TRUST.1).

  • verify_anchor

    Verify the cryptographic integrity of an SWT3 witness anchor.

  • witness_adapter_stack

    Witness active LoRA/QLoRA/PEFT adapter stack (AI-MDL.6).

  • witness_adr_baseline

    Witness baseline consumption (ADR-BASE.1). FERC Order 2222. Evidence only.

  • witness_adr_carbon

    Witness carbon credit / REC (ADR-CARBON.1). EU CBAM, EU RED III. Evidence only.

  • witness_adr_curtailment

    Witness curtailment verification (ADR-CURT.1). FERC Order 2222. Evidence only.

  • witness_adr_event

    Witness demand response event (ADR-EVENT.1). FERC Order 2222. Evidence only.

  • witness_adr_grid_signal

    Witness grid signal correlation (ADR-GRID.1). FERC Order 2222, NERC BAL-001. Evidence only.

  • witness_adr_settlement

    Witness settlement data (ADR-SETTLE.1). FERC Order 2222. Evidence only.

  • witness_agent_card_discovery

    Witness A2A Agent Card discovery event (AI-A2A.2).

  • witness_agent_handoff

    Witness inter-agent handoff (AI-ORCH.2). Records when one agent delegates work to another with permission boundary changes. Evidence only.

  • witness_anchor_density

    Witness anchor density evaluation (AI-DENSITY.1).

  • witness_authorization

    Witness an authorization decision as an AI-ACC.1 anchor.

  • witness_consent

    Witness data subject consent or lawful basis documentation (AI-CONSENT.1).

  • witness_context_chain

    Witness A2A context chain linking event (AI-A2A.3).

  • witness_context_window

    Witness context window management (AI-CTX.1). Records when a harness truncates, summarizes, or evicts context tokens. Evidence only.

  • witness_data_provenance

    Witness training data governance diligence (AI-DATA.1).

  • witness_delegation_boundary

    Witness delegation boundary evaluation (AI-DEL.2).

  • witness_delegation_tree

    Witness a hierarchical delegation tree grant (AI-DEL.1).

  • witness_distillation

    Witness knowledge distillation provenance (AI-DIST.1).

  • witness_dpp_chargewrite action

    Witness charge/discharge cycle (DPP-CHRG.1). EU Battery Reg Art. 14(1). Evidence only.

  • witness_dpp_degradation

    Witness battery degradation (DPP-DEGRAD.1). EU Battery Reg Art. 14(1). Evidence only.

  • witness_dpp_eol

    Witness end-of-life handoff (DPP-EOL.1). EU Battery Reg Art. 59. Evidence only.

  • witness_dpp_soh

    Witness battery State of Health (DPP-SOH.1). EU Battery Reg Art. 14(1). Evidence only.

  • witness_elicitation

    Witness MCP elicitation consent (AI-MCP.6).

  • witness_eval_gate

    Witness eval gate decision (AI-GATE.1). Records pass/fail deployment gating based on eval results. Auto-computes gate score when not provided. Evidence only.

  • witness_guardrail

    Witness guardrail implementation and activation (AI-GRD.1).

  • witness_hbom_inventory

    Witness hardware inventory attestation (HBOM-INV.1). EU CRA Art. 10(9). Evidence only.

  • witness_hbom_lifecycle

    Witness component lifecycle event (HBOM-LIFE.1). EU Battery Reg Art. 77. Evidence only.

  • witness_hbom_pue

    Witness PUE attestation (HBOM-PUE.1). EU EED Art. 12, ISO 30134-2. Evidence only.

  • witness_hbom_supply

    Witness hardware supply chain provenance (HBOM-SUPPLY.1). EU CRA Art. 10(9). Evidence only.

  • witness_hbom_thermal

    Witness thermal profile (HBOM-THERM.1). ASHRAE TC 9.9. Evidence only.

  • witness_hbom_water

    Witness water consumption (HBOM-WATER.1). CSRD ESRS-E3. Evidence only.

  • witness_human_review

    Witness that human review occurred on AI-generated output (AI-HITL.1).

  • witness_incident

    Witness incident detection and reporting (AI-INCIDENT.1).

  • witness_inference

    Mint a cryptographic SWT3 witness anchor for an AI inference.

  • witness_mcp_security

    Witness MCP security posture evaluation (AI-MCP.1).

  • witness_model_integrity

    Witness model weight file integrity (AI-MDL.5).

  • witness_model_provenance

    Witness model provenance chain (AI-PROV.1).

  • witness_nhi_delegation

    Witness agent-to-agent credential delegation (NHI-AGENT.1). Records delegation chain depth and participants. Evidence only.

  • witness_nhi_expiration

    Witness credential expiration event (NHI-EXPIRE.1). Records when a credential expires, whether renewal is possible, and grace period. Evidence only.

  • witness_nhi_lifecycle

    Witness credential lifecycle event (NHI-CYCLE.1). Records issuance, activation, suspension, expiration, or revocation. Evidence only.

  • witness_nhi_privilege_change

    Witness privilege change on a credential (NHI-PRIV.1). Records scope escalation or de-escalation. Evidence only.

  • witness_nhi_revocation

    Witness credential revocation (NHI-REVOKE.1). Records which credential was revoked, why, and whether it cascades. Evidence only.

  • witness_nhi_rotation

    Witness credential rotation (NHI-ROTATE.1). Records old-to-new credential swap with reason. Evidence only.

  • witness_nhi_scope

    Witness credential scope attestation (NHI-SCOPE.1). Creates an audit trail of reported credential permissions. SWT3 is not the credential authority -- it witnesses what the IdP reports. Evidence only.

  • witness_oauth_token_binding

    Witness OAuth token lifecycle event for MCP servers (AI-MCP.5).

  • witness_orchestration_topology

    Witness orchestration topology selection (AI-ORCH.1). Records when a harness selects sequential, parallel, hierarchical, hybrid, or mesh routing for a multi-agent task. Evidence only.

  • witness_output_filter

    Witness output content safety classification result (AI-GRD.2).

  • witness_rag_context

    Witness RAG (Retrieval-Augmented Generation) context retrieval (AI-RAG.1).

  • witness_resource_consumption

    Witness cumulative resource consumption (AI-COST.1).

  • witness_runtime_containment

    Witness runtime containment attestation (AI-SHELL.1). Records that a sandboxed runtime (OpenShell, gVisor, Kata, Firecracker, WASM) enforced its containment policy. Duck-typed: no runtime SDK dependency. Evidence only.

  • witness_sandbox_enforcement

    Witness sandbox enforcement attestation (AI-SAND.1). Records the harness's own report of tool restriction compliance. Cross-reference AI-TOOL.1 for independent verification. Evidence only.

  • witness_server_auth

    Witness MCP server authentication attestation (AI-MCP.3).

  • witness_server_discovery

    Witness MCP server discovery attestation (AI-MCP.4).

  • witness_task_lifecycle

    Witness A2A task delegation lifecycle event (AI-A2A.1).

  • witness_tool_integrity

    Witness MCP tool integrity attestation (AI-MCP.2).

  • witness_trajectory

    Witness a safety-critical trajectory decision from a VLA or autonomous planning model (AI-MOB.6).

Public scan report

scanner v0.1.10 · 2026-10-29 · same rubric, same numbers if you re-run it

no findings
  • Code scan105 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 qualitystatic API keys via environment variables6/15
  • Maintenancelast push 2 days ago15/15
  • Maintainer identitynamespace and repository owner differ; website matches verified namespace6/10
Overall 80/100. Components that don't apply are left out of the denominator. Any critical finding is an F.RubricAppeal a findingJSON

How SWT3 AI Witness compares

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

Safety score
80/100
higher than 60% of 48,243 servers
Tools
75
the typical server has 7
Tools that change things
1%
typical: 0%. Fewer write tools means less an agent can break
Downloads a week
670
more than 91% of 12,306 servers with a package

What its tools let an agent do: 72 read · 1 change · 2 run or deploy

Release history: 3 versions in the registry since 2026-09-01, latest 2026-10-06, typically one every 21 days.

Install SWT3 AI Witness in Claude Code, Cursor or VS Code

Runs npx -y @tenova/swt3-mcp on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add swt3-witness -- npx -y @tenova/swt3-mcp
Add to Cursor

What the publisher says

From the SWT3 AI Witness repository's README, as published. We do not edit it. Read it on GitHub

SWT3 - Sovereign Witness Protocol for AI

Don't audit the agent's thoughts. Audit the agent's actions.

The Problem

AI agents are making production decisions: approving loans, triaging patients, managing infrastructure, writing code. In 2026, 65% of firms reported AI agent security incidents. Only 14.4% of agents go live with full security approval. When something goes wrong, there is no tamper-proof audit trail. Logs are mutable. Metrics are averaged. Nobody can prove what the agent actually did.

GPAI transparency obligations are enforceable now. EU AI Act high-risk enforcement begins December 2, 2027. NIST AI RMF, SR 11-7, and CMMC impose similar obligations. 72% of enterprises believe they have AI governance but lack actual control. Most teams have nothing but dashboards and hope.

The Protocol

SWT3 (Sovereign Witness Traceability) is a deterministic witness protocol for AI systems. It intercepts AI actions, hashes the evidence, and anchors cryptographic proof to an immutable ledger. Your code gets the full response. The auditor gets tamper-proof evidence. Raw prompts and responses never leave your infrastructure.

  • Deterministic, not probabilistic. The witness engine uses fixed logic, not AI, to evaluate compliance.
  • Zero data retention. Configurable clearing levels strip sensitive content before it leaves your environment.
  • Framework-mapped. Every anchor maps to EU AI Act articles, NIST AI RMF functions, and federal controls.

Try It (10 Seconds, No Account)

Python

pip install swt3-ai
python -m swt3_ai.demo

TypeScript

npm install @tenova/swt3-ai
npx swt3-demo

No API keys. No account. No network calls. You will see the full witnessing pipeline run locally.

Three Lines to Production

from swt3_ai import Witness
from openai import OpenAI

witness = Witness(endpoint="https://sovereign.tenova.io", api_key="axm_live_...", tenant_id="YOUR_TENANT")
client = witness.wrap(OpenAI())

# Every inference is now witnessed. Your code does not change.
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Summarize this contract"}],
)

Works with OpenAI, Anthropic, AWS Bedrock, Vercel AI SDK, LangChain, LiteLLM (100+ providers), and any OpenAI-compatible endpoint (vLLM, Ollama, Azure OpenAI).

K8s Hardware Attestation

AI compliance doesn't stop at the model layer. Regulators want to know what hardware ran the inference, whether it was authorized, and whether the compute environment changed between audit periods. Today, no platform provides cryptographic proof of which silicon processed which workload. SWT3 does.

Deploy a DaemonSet that discovers accelerator hardware on every node and mints AI-HW.1 attestation anchors. Zero application code changes. One Helm install.

helm install swt3-witness oci://ghcr.io/tenova-labs/charts/swt3-witness \
  --set config.endpoint="https://sovereign.tenova.io" \
  --set config.apiKey="axm_live_..." \
  --set config.tenantId="YOUR_TENANT"

The DaemonSet auto-discovers 6 accelerator types per node:

Shortened. The full README is on GitHub.

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

SWT3 AI Witness: common questions

Is SWT3 AI Witness MCP server safe?
Mostly: it is graded B (80/100). Read the SWT3 AI Witness safety report
How do I install SWT3 AI Witness?
It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does SWT3 AI Witness need an API key?
Yes. The registry entry asks for SWT3_API_KEY, SWT3_SIGNING_KEY.
Is SWT3 AI Witness maintained?
The last commit was 9 days ago (2026-10-27). The latest release is v0.7.5.
How does SWT3 AI Witness compare with other MCP servers?
safety score 80/100 (higher than 60% of 48,243 servers); tools 75 (the typical server has 7); tools that change things 1% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.