mnox

by Matt Noxon

Agentic Implementation Optimizer — audit, build, or extend agentic AI implementations for reliability, cost, and architecture.

by Matt Noxon

Meta-plugin that installs every mnox-ai agentic-review skill at once via plugin dependencies. Standalone utility plugins (session-tracker, config-chunks, create-skill, bash-gate) are installed separately. Excludes foundry, which is decommissioned for public use.

by Matt Noxon

Browser-driven job application copilot — locate matching postings, score fit, tailor resume/cover letter/screening answers per job to score well on real evaluator engines (Workday HiredScore, Taleo, iCIMS, Ashby, LinkedIn, LLM screeners) with a deterministic keyword-coverage check, drive the application form with a human confirmation gate before every submit, and track the pipeline locally.

by Matt Noxon

A PreToolUse Bash gate that auto-allows provably-safe commands (so you stop approving the same harmless `mkdir`/`rm` of a build artifact) while still deferring anything dangerous. Deterministic allow-classes by default; an optional LLM arbiter handles the long tail. Ships safe-by-default (no path is auto-allowed until you opt in) and a /bash-gate-add skill to extend it when a prompt slips through.

by Matt Noxon

Audit a repo, ADR/PRD, IaC, or live cloud state against SOC 2, HIPAA, or PCI-DSS with structured, evidence-backed findings — or assess organizational readiness for a SOC 2 examination (governance, evidence gaps, phased remediation).

by Matt Noxon

A package manager for agent-instruction guidance — publishes versioned, scored guidance 'chunks', reconciles them into a single bundle, and wires it into CLAUDE.md (@import) and/or AGENTS.md (inlined). Includes /ai-setup onboarding and a chunk-review scoring rubric.

by Matt Noxon

Author a Claude Code (or any host's) Agent Skill end-to-end — scaffold, write against a portable authoring contract, structurally validate, and design-score — in one guided lifecycle. Spec-pure by default; Claude-only affordances are opt-in.

by Matt Noxon

Generate an adaptive learning curriculum for any topic, with an append-only assessment loop that tunes future modules to the learner.

by Matt Noxon

Pre-public open-source readiness audit — secrets/PII, licensing, community-health files, tests/CI, deps — scored with fix commands.

by Matt Noxon

Diagnose slow database queries across Postgres, the ORM layer (N+1), and Snowflake/dbt — routes to five modes (known query, locate offender, live incident, proactive audit, regression) and produces a root-caused, verified fix.

by Matt Noxon

DECOMMISSIONED for public use. The original autonomous-fix loop (Worker/Integrator over isolated git worktrees) has been retired; the skill is now a compatibility shim that routes into the maintainer's private Sven Unit runtime, not distributed in this repo. Not functional for other marketplace installs.

by Matt Noxon

Audit a repo, design doc, IaC, or data-flow map against GDPR — lawful basis, consent, data subject rights, retention/erasure, transfers, processors, privacy by design — with article-cited, evidence-backed findings, or assess organizational GDPR readiness. Analysis aid, not legal advice.

by Matt Noxon

Audit a knowledge graph or ontology for structural health across seven axes — orthogonality, granularity, taxonomic hygiene, identity & rigidity, relationship semantics, competency questions, and inference safety — and flag defects before they corrupt inference.

by Matt Noxon

Audit a retrieval / vector-index pipeline for structural health across seven layered axes — eval foundation, corpus & chunking, embedding geometry, index & ANN fidelity, retrieval composition, rank fusion, and reranking — and flag the defects that silently return wrong or missing context before they reach the generator.

by Matt Noxon

Review database schemas and in-code data structures for correctness, integrity, performance, scalability, and design quality.

by Matt Noxon

Local MCP server that tracks your Claude Code / Cursor / Codex agent sessions — list, search, label, and inspect changes across every session. Lexical (FTS5) search by default; semantic search is opt-in.

by Matt Noxon

Clean-room legacy reimplementation — distill to requirements, rebuild greenfield behind a context firewall, audit each crossing for leaked legacy structure (taint/provenance), verify parity, optionally cut over.

by Matt Noxon

Review Agent Skills, hooks, AGENTS.md/CLAUDE.md files, scripts, MCP configs, or workflow configurations for design flaws, unclosed loops, stale references, and risks.