Self-improving agent rig — codifies product and engineering discipline into composable workflows from brainstorm to ship. Compounds knowledge, generates domain agents, and monitors its own docs. Orchestrates Claude, Codex, and GPT-5.2 Pro through 4 agents, 54 commands, 16 skills, 1 MCP servers. Companions: interphase, interline, interflux, interpath, interwatch, interslack, interform, intercraft, interdev, interpeer, intertest.
28 skills · 65 commands · 7 agents · hooks
Multi-agent review and research with scored triage, content slicing, and intermediate finding sharing. 17 agents (12 review + 5 research), 4 commands, 1 skill, 1 MCP server. Domain detection via intersense, knowledge via interknow.
3 skills · 8 commands · 8 agents · 2 MCP servers · hooks
Agent performance profiler and routing optimizer. Collects evidence about flux-drive agent accuracy, proposes routing overrides for underperforming agents, and monitors canary periods. Companion plugin for Clavain.
18 commands · hooks
Bidirectional Notion sync for Claude Code. Three-way merge conflict resolution, beads issue tracking, and 21 MCP tools for project lifecycle and sync control.
4 skills · 3 commands · 7 agents · 1 MCP server · hooks
Token-efficient code reconnaissance for LLMs. Autonomous skills save 48-85% tokens via diff-context, semantic search, structural patterns, and symbol analysis. Includes MCP server for direct tool integration.
5 skills · 7 agents · 1 MCP server
Voice engine with a closed loop: generate and edit prose in your voice, then verify it against your measured stylometric fingerprint. Register-aware profile, LLMism linter, exemplar retrieval, critique-revise apply. Successor to interfluence and intervoice.
9 skills · 1 command · 1 agent · hooks
Product artifact generator — roadmaps, PRDs, vision docs, changelogs, and status reports from beads state, brainstorms, and project context. Companion plugin for Clavain.
1 skill · 8 commands
Interactive ecosystem diagram — scans Interverse monorepo and generates a D3.js force graph showing all plugins, skills, MCP tools, hooks, and their relationships.
1 skill · 7 agents
Ambient discovery and research engine — scans arXiv, HN, GitHub, Anthropic docs, Exa for new capabilities. MCP server with embedding-based recommendation engine, confidence-tiered output pipeline, and learned interest profiles.
6 skills · 1 MCP server · hooks
Multi-agent file coordination — reserve files before editing, detect conflicts, exchange messages between agents. MCP server wrapping intermute with hooks, commands, skills, and git pre-commit enforcement. Companion plugin for Clavain.
2 skills · 4 commands · 1 MCP server · hooks
Doc freshness and correctness monitoring — auto-discovers watchable docs, detects drift via 17 signal types, scores confidence, dispatches generators for auto-refresh, and runs stranger-perspective correctness audits with cross-document consistency checking. Companion plugin for Clavain.
1 skill · 5 commands · hooks
Deep research plugin — content extraction and research orchestration via MCP tools.
1 skill · 1 command · 3 agents · 1 MCP server
Recursive AGENTS.md generator with structural auto-fix, CLAUDE.md harmonization, incremental updates, diff previews, and smart monorepo scoping. Cross-AI compatible.
1 skill · 1 command · 4 agents
TUI automation and visual testing - Playwright for terminal applications. Spawn, interact with, and screenshot TUI apps via MCP.
1 skill · 4 agents · 1 MCP server
Agent-as-operator runtime diagnostics — teaches agents to observe and control running applications via diagnostic HTTP servers and CLI tools. 3 skills, 1 reviewer agent, 3 PostToolUse hooks, Rust templates.
4 skills · hooks
Token efficiency benchmarking for agent workflows — PostToolUse:Task hook captures real-time events to SQLite, SessionEnd hook backfills token data from JSONL, report shows percentiles and decision gate.
4 skills · hooks
Knowledge compounding — durable pattern repository with provenance tracking, temporal decay, and semantic retrieval via qmd MCP server.
2 skills · 1 MCP server · hooks
Autonomous experiment loop with mutation provenance tracking — init, run, log experiments with JSONL persistence, git isolation, SQLite mutation store, and ic events bridge.
2 skills · 1 MCP server · hooks
Philosophy observer — detects latent design patterns and philosophy drift from decision artifacts, proposes PHILOSOPHY.md updates.
1 skill · 3 commands
Multi-agent synthesis engine — collects findings from parallel review/research agents, deduplicates, writes structured verdicts, produces compact summaries.
3 agents · hooks
Dynamic statusline for Claude Code — shows active beads with priority and title, workflow phase, coordination status, and Codex dispatch state. Integrates with Clavain, interphase, interlock, and beads.
2 commands · hooks
Memory synthesis — graduates stable auto-memory facts to AGENTS.md/CLAUDE.md with stability detection, deduplication, interactive approval, and atomic pruning.
2 skills · hooks
Agent activity visibility — tmux monitoring, activity feeds, health detection. Enriches intermute with live context. 7 MCP tools for cross-agent observability.
1 skill · 1 MCP server · hooks
Plugin development toolkit — lifecycle management, structure validation, and troubleshooting for Claude Code plugins. Extracted from interdev.
3 skills
Shared embedding and search infrastructure — nomic-embed-text-v1.5 embeddings with persistent SQLite vector store, and Exa web search. MCP server for semantic indexing and querying.
2 skills · 1 MCP server
Engineering quality disciplines — systematic debugging (4-phase root-cause investigation), test-driven development (strict RED-GREEN-REFACTOR), and verification gates (iron-law evidence checks).
3 skills
Tool usage analytics for Claude Code and Codex CLI. Tracks tool patterns via hooks, detects inefficiencies, and offers opt-in community comparison with anonymized data.
2 skills · hooks
Code quality guards — syntax validation for 8 languages and auto-formatting via ruff/shfmt/gofmt/jq/prettier. Session monitoring extracted to interpulse.
1 skill · hooks
Agent-native architecture patterns — design, review, and audit for agent-first applications.
1 skill · 1 command
Deploy monitoring and auto-fix — watches Vercel deployments after push, diagnoses build failures, fixes code, re-pushes until success.
1 skill · hooks
Developer tooling for Claude Code — MCP CLI interaction and Claude Code reference. Skill authoring moved to interskill, plugin development moved to interplug.
2 skills
Cross-repo institutional knowledge index — indexes solution docs across the Interverse monorepo, enables unified search, and audits reflect coverage.
1 skill · hooks
Project-level code mapping MCP server: project registry, call graphs, impact analysis, agent overlay. Go + Python bridge.
1 skill · 1 MCP server
Cross-repo matrix evaluation harness — run Skaffen against unfamiliar codebases, classify outcomes, track failure modes.
1 skill · 1 MCP server
Agent and agency naming for legible orchestration — deterministic, memorable identities across all surfaces. 4 themes: Culture (Banks-style), Sylveste (diplomatic), NATO (phonetic), custom.
2 commands
Cross-AI peer review with 4 escalation modes — quick (Claude/Codex), deep (Oracle/GPT), council (multi-model consensus), mine (disagreement extraction into tests and specs).
1 skill · 1 command
Phase tracking, gate validation, and work discovery for the Beads issue tracker. Companion plugin for Clavain — adds lifecycle state management on top of the core beads plugin.
1 skill · hooks
Session context monitoring — pressure tracking with time-decay, token estimation, and threshold warnings. Extracted from intercheck.
1 skill · hooks
Culture ship names as Claude Code spinner verbs. 237 curated names -- 56 canonical Banks originals marked with *asterisks*, 181 generated Banksian originals. Toggle canonical/generated sets via config. Includes v6 generator prompt for creating more.
1 command · hooks
Skill authoring toolkit — unified skill creation with TDD-adapted testing and audit. Extracted from interdev.
2 skills
Cross-session semantic cache for Claude Code. Content-addressed blob storage, per-project manifests, and session tracking — reduces cold start time and eliminates redundant file reads across sessions.
1 MCP server
Local MLX-LM inference server for Apple Silicon. Custom serving layer with priority queuing and thermal-aware scheduling.
1 MCP server
Design patterns and visual quality for Claude Code — distinctive, production-grade interfaces.
1 skill
Deterministic Skeleton with LLM Islands — token-efficient document generation via template-then-fill, where scripts render predictable sections and LLM subagents fill semantic placeholders.
1 skill
Proof-oriented agent loops for multi-project solo-lab work. Selects demo, release, repo hygiene, installer, eval, runtime, and handoff loops with explicit evidence and stop conditions.
1 skill
Hegelian dialectic reasoning — Electric Monk subagents for structured contradiction analysis and synthesis
1 skill
Work prioritization and next-task analysis — tradeoff-aware recommendations from beads state and project context.
1 skill
Snapshot-backed model/benchmark ranking MCP server for AgMoDB
1 MCP server
Documentation quality toolkit — enforces CLAUDE.md/AGENTS.md boundaries, applies progressive disclosure, deduplicates across doc hierarchy. 1 skill with audit, refactor, and consolidate modes.
1 skill
Automated UI/UX design analysis — extracts W3C DTCG tokens, component inventory, and layout analysis from any URL using Dembrandt + Playwright MCP + Claude vision.
1 skill
Slack integration for Claude Code — send messages, read channels, test integrations.
1 skill
Cross-module integration gap tracer — traces data flows from shipped features to find unverified consumer edges.
1 skill
Project hierarchy management — filesystem discovery, parent-child relationships, tagging, and layout orchestration. Extracted from interkasten.
1 skill
Agent trust scoring — severity-weighted time-decayed reputation tracking. Records finding acceptance/rejection, computes blended project/global trust scores, identifies suppression candidates.
1 command
CUJ design system: discover → teardown → provoke → capture → derive → consume; evidence-first schema; agent-consumable specs for coding agents to build from with minimal drift.
Browser automation and competitive research for AI agents. UX teardowns, docs crawls, cross-cutting synthesis, and CUJ generation — the full research-to-design pipeline via agent-browser CLI.
Friction journal: papercut CLI + /papercuts review — log small friction in the moment, cluster it later
Safe plugin publishing — bumps all version locations, validates sync, commits and pushes with confirmation.
No description.
Feature-level success metric tracking for Sylveste