Kineti MCP server

by therawlogs·io.github.therawlogs/kineti·v0.3.4

No reviews yet

Kineti: spend cap, undo, and test proof for AI coding agents

A89/100grade A

Write a review Install

What users say

No reviews yet

Be the first

Safety scan

A89/100

full report

Adoption

New

0 stars

Reviews

Write one

Nobody has reviewed Kineti yet.

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

Kineti tools (14, 2 write)

write = sends, deletes, buys or posts

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

kineti_egress_record
Record an outbound network request in the tamper-evident cryptographic egress ledger.
kineti_epistemic_eval
Evaluate a candidate action against the 360º Human Model (Rule-Exception hierarchies, Epistemic certainty, and Safety ceilings).
kineti_evidence_check
Check if an existing verification proof is FRESH or STALE against current workspace code.
kineti_evidence_recordwrite action
Execute a verification test command and record cryptographic SHA-256 fingerprint proof. Takes a command array against an allowlist (bun test, pytest, npm test, project verify). Shell needs human --allow-shell plus TTY.
kineti_lock_goal
Lock the immutable root goal for the current task run. Fails if goal is already locked.
kineti_saga_push
Register an undo action on the LIFO rollback stack before executing a state mutation. Takes a command array against an allowlist. Shell needs human --allow-shell plus TTY.
kineti_saga_rollback
Rollback uncommitted actions on the LIFO saga stack. Human-only: needs a TTY y/N or explicit --yes; MCP calls fail closed with instructions.
kineti_set_gate
Set status for a pipeline quality gate ('spec', 'feasibility', 'security', 'audit').
kineti_set_stagewrite action
Update the active stage or task in the Kineti OS runtime (e.g. 'build', 'spec', 'bugfix', 'refactor', or numbers 1-13).
kineti_spend_record
Record LLM token consumption and enforce stage and task financial circuit breakers.
kineti_spend_status
Check current spend across stages, total task spend, and circuit breaker trip status.
kineti_status
Retrieve current Kineti harness status: active project, locked root goal, current stage, and verification gates.
kineti_talk
Talk to Kineti in plain words. No skill or command names needed. Handles spending, undo, proof, status, yes/no, on/off, and routes tasks. Always returns plain words plus numbered choices.
kineti_verify_gate_status
Inspect pre-flight verify gate status and current repository trust.

Public scan report

scanner v0.1.9 · 2026-09-25 · same rubric, same numbers if you re-run it

no findings

Code scan53 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 0 days ago15/15
Maintainer identityregistry namespace matches repository owner6/10

Overall 89/100. Components that don't apply are left out of the denominator. Any critical finding is an F.Rubric Appeal a finding JSON

How Kineti compares

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

Safety score: 89/100; higher than 80% of 48,255 servers
Tools: 14; the typical server has 7
Tools that change things: 14%; typical: 0%. Fewer write tools means less an agent can break

What its tools let an agent do: 12 read · 2 change

Install Kineti in Claude Code, Cursor or VS Code

Runs npx -y kineti on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add kineti -- npx -y kineti

Add to Cursor

What the publisher says

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

Kineti OS

The Context Integrity Layer (Context Integrity Protocol / CIP) & Autonomous Nervous System

Kineti OS is a dual-stack agent verification harness and Context Integrity Layer (Context Integrity Protocol / CIP) for AI coding agents. Its wedge is spend ceilings, transactional undo (SAGA), project-scoped memory, and cryptographic evidence receipts. It combines a Native Rust Nervous System (core-native/) with a deterministic TypeScript Governance Control Plane (bin/, src/). A 5-part foundational research series is forthcoming at therawlogs.com. Consumer messaging connectors (WhatsApp/iMessage) and payment tooling in this repo are experimental.

Key Pillars

1. The 360º Human Model & Epistemic Engine

Personalized agents must understand the full human context without hallucinating, overriding user statements, or quietly mutating goals when third-party systems push back:

Multi-Scope Context Isolation: Strict context isolation across Global, Domain (Health, Work, Finance, Schedule, Taste), and Relationship (person-to-person) scopes. Scoped facts never leak into generic or un-scoped queries.
Epistemic Certainty Tiers: Enforces DirectlyKnown (explicit user ground truth) > ObservedPattern (behavioral patterns) > Inferred (hypotheses). Machine inferences are strictly blocked from overwriting explicit user statements.
Rule-Exception Hierarchies: Resolves complex user lifestyles unambiguously: BaselineRule (e.g. Vegetarian) $\to$ PermittedException (e.g. Eats eggs) $\to$ Preference (e.g. Prefers low dairy) $\to$ SafetyCeiling (e.g. Peanut allergy).
Verbatim Root Goal & Anti-Drift Engine: Anchors autonomous task chains to the exact, unmodified words uttered by the user and their explicit definition of "Done". Every step is inspected against the original ask, eliminating the multi-step "telephone game". Intermediate steps and tools are treated as expendable scaffolding.
Friction Triage Ladder ("Clean No over Dirty Yes"): Triages real-world obstacles through 3 levels:

Noise: Transient blips auto-retry with exponential backoff.
Broken Surface: Broken websites or portals silently reroute to alternatives.
Real Constraint: Hard third-party refusals escalate immediately with a clean impossibility report. Sunk costs are written off ($0 sunk-cost fallacy), and quiet compromises (such as accepting budget overruns) are strictly blocked.

Commitment-Time Verification: Re-checks perishable facts (fares, seat availability, stock levels, auth tokens) at the exact millisecond of external or financial commit, never trusting cached plan snapshots.
Asymmetric Gap-Filling: Cheap, reversible gaps are filled automatically and disclosed in audit evidence; expensive or irreversible gaps halt execution to ask the user.
Reputation Gating & Ingress Defense: Outbound communication is treated as a non-regenerating resource (knowing an identity does not equal permission to contact). Incoming external messages and webhooks arrive as untrusted data, never instructions.

Shortened. The full README is on GitHub.

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

Kineti: common questions

Is Kineti MCP server safe?: Yes, by our scan: it is graded A (89/100). Read the Kineti safety report
How do I install Kineti?: It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does Kineti need an API key?: Not as far as the registry entry and our scan can tell: no credentials are declared or required.
Is Kineti maintained?: The last commit was 40 days ago (2026-09-25). The latest release is v0.3.4.
How does Kineti compare with other MCP servers?: safety score 89/100 (higher than 80% of 48,255 servers); tools 14 (the typical server has 7); tools that change things 14% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.