Labs Catalog
ProjectPulse
// MARKETPLACEForge Orchestrator
NXTG-Forge orchestrator: an MCP server for Claude Code with task orchestration, file locking, knowledge capture and drift detection. Rust. Alpha (source-available, FSL-1.1-ALv2).
Forge Plugin
NXTG-Forge plugin for Claude Code: governance, 33 agents, 24 commands, 27 skills, 27 hook registrations. Alpha (MIT).
Forge Ui
NXTG-Forge dashboard: Infinity Terminal and a view of the project's governance state. React. Public preview (source-available, FSL-1.1-ALv2).
Nxtg Forge
Superseded by forge-plugin, forge-orchestrator, and forge-ui
SynApps V0.4.0
SynApps is a web-based visual platform for modular AI agents called Snaplets. Its mission is to let indie creators build autonomous AI snaplets like LEGO blocks – each snaplet is a small agent with a specialized skill (e.g. Writer, Memory, Artist).
Faultline Pro
FM-agnostic AI Trust & Safety scanner — multi-provider grounded claim verification (Apache-2.0 CLI + optional hosted tier)
Jev Calibration Eval
Pre-registered calibration study of Jev vs a local model and a frontier model (S1 data + harness)
Nxtg Atlas
Portfolio intelligence for AI engineering teams. Scan repos, score health, find cross-project patterns.
.Github
No description available
Forge
Forge development workspace. See forge-plugin, forge-orchestrator, and forge-ui for the product repos.
Homebrew Tap
Homebrew tap for NXTG-AI tools (forge-orchestrator)
Forge Done Fixture 2
Deliberately seeded, fixture 2 (harder): a change marked done whose test run says OK, broken in a way its diff does not show. Pre-registered: does Forge beat a plain review?
Forge Done Fixture
Deliberately seeded: a change marked done whose tests pass but miss a defect. Does Forge catch it? Pre-registered runs, raw outputs, red and green tags.
Faultline Action
Forensic AI output verification for your CI pipeline — detects hallucination, manipulation, and EU AI Act violations
Judge Fraud Gates
Eval-native gate family that catches semantic LLM-judge fraud standard CI is blind to — 27 labeled judge harnesses, measured confusion matrix, one-command repro (MIT)
Cosmux
CoS-aware tmux pod manager — declarative YAML, lifecycle hooks (Rust)
Voice Jib Jab
NextGen AI Voice Orchestrator with lane-based architecture - Browser-first speech-to-speech assistant
Atlas Action
Atlas — repo health audit GitHub Action. Posts stack/health/recommendations as PR comment. Powered by nxtg-atlas.