interagency-marketplace

Self-improving agent rig: codifies product and engineering discipline into composable workflows from brainstorm to ship. Compounds knowledge, generates domain agents, monitors its own docs, and surfaces conservative update drift. Orchestrates Claude, Codex, and Oracle through 6 agents, 57 commands, 20 skills, 0 MCP servers. Factory substrate: CXDB turn DAG, scenario bank with satisfaction scoring, evidence pipeline, agent capability policies. Companions: interspect, interphase, interline, interflux, interpath, interwatch, interslack, interform, intercraft, interdev, interpeer, intertest.

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.

Interactive ecosystem diagram — scans Interverse monorepo and generates a D3.js force graph showing all plugins, skills, MCP tools, hooks, and their relationships.

Code quality guards and session health monitoring

Dynamic statusline for Claude Code — shows workflow phase, bead context, and Codex dispatch state. Integrates with Clavain and interphase.

Agent-native architecture patterns — design, review, and audit for agent-first applications.

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.

Deep research plugin — content extraction and research orchestration via MCP tools.

Deploy monitoring and auto-fix — watches Vercel deployments after push, diagnoses build failures from logs, fixes code, and re-pushes until the deploy succeeds.

Developer tooling for Claude Code — MCP CLI interaction and Claude Code reference.

Recursive AGENTS.md generator with integrated Oracle critique, CLAUDE.md harmonization, incremental updates, diff previews, and smart monorepo scoping. Cross-AI compatible.

Local MLX-LM inference server for Apple Silicon. Custom serving layer with priority queuing, thermal-aware scheduling, and experiment hooks for Sylveste/Clavain.

[DEPRECATED — use intervox] Analyze your writing style and adapt Claude's output to sound like you. Replaced by intervox (via the interim intervoice), which adds a measured stylometric fingerprint, LLMism linter, and verified closed-loop apply on top of the global multi-register profile.

Multi-agent review and research with scored triage, domain detection, content slicing, intermediate finding sharing, and knowledge injection. 17 agents (12 review + 5 research), 8 commands (incl. flux-melange — a goal-seeking adaptive review loop with lens fusion and novelty/risk/taste scoring), 3 skills (flux-engine, flux-review-engine, flux-melange-engine), 2 MCP servers (exa, openrouter-dispatch). Companion plugin for Clavain.

Design patterns and visual quality for Claude Code — distinctive, production-grade interfaces.

Agent-as-operator runtime diagnostics — teaches agents to observe and control running applications via diagnostic HTTP servers and CLI tools.

Ambient discovery and research engine. Continuously scans arXiv, Hacker News, GitHub, and Anthropic docs for new capabilities, workflows, and tools relevant to your engineering ecosystem. Creates beads with briefings and draft plans based on a learned recommendation model.

Living bridge between your projects folder and Notion — bidirectional sync with adaptive AI documentation and pagent workflow automation

Knowledge compounding — durable pattern repository with provenance tracking, temporal decay, and semantic retrieval via qmd.

Autonomous experiment loop with mutation provenance tracking — init, run, log experiments with JSONL persistence, git isolation, SQLite mutation store, and ic events bridge.

Cross-repo institutional knowledge index — indexes solution docs across the Interverse monorepo, enables unified search, and audits reflect coverage.

Deterministic Skeleton with LLM Islands — token-efficient document generation via template-then-fill.

288 FLUX cognitive lenses for structured thinking — MCP server with lens search, thinking mode workflows, belief statement generation, quality evaluation, and solution synthesis.

MCP server for intermute file reservation and agent coordination. 12 tools: reserve, release, conflict check, messaging, agent listing, negotiation, escalation. Companion plugin for Clavain.

Proof-oriented agent loops for multi-project solo-lab work.

Philosophy observer — detects latent design patterns and philosophy drift from decision artifacts, proposes PHILOSOPHY.md updates.

Project-level code mapping: project registry, call graphs, architecture analysis, agent overlay, cross-project dependencies, pattern detection, live changes. MCP server with 9 tools.

Memory synthesis — graduates stable auto-memory facts to AGENTS.md/CLAUDE.md

Cross-repo matrix evaluation harness — run Skaffen against unfamiliar codebases, classify outcomes, track failure patterns.

Hegelian dialectic reasoning — Electric Monk subagents for structured contradiction analysis and synthesis

Agent activity visibility -- tmux monitoring, activity feeds, health detection. Enriches intermute with live context.

Agent and agency naming for legible orchestration — memorable, deterministic identities across all surfaces.

Work prioritization and next-task analysis for Claude Code — tradeoff-aware recommendations from project state.

Product artifact generator — roadmaps, PRDs, vision docs, changelogs, CUJs, and status reports from beads, brainstorms, and project state. Companion plugin for Clavain.

Cross-AI peer review — quick (Claude↔Codex), deep (Oracle), council (multi-model), mine (disagreement extraction).

Documentation quality toolkit — enforces CLAUDE.md/AGENTS.md boundaries, applies progressive disclosure, deduplicates across doc hierarchy.

Plugin development toolkit — lifecycle management, structure validation, and troubleshooting for Claude Code plugins

Safe plugin publishing — bumps all version locations, validates sync, commits and pushes with confirmation.

Session context monitoring — pressure tracking, token estimation, threshold warnings

Snapshot-backed model/benchmark ranking MCP server for AgMoDB

Shared embedding and search infrastructure for the Interverse ecosystem — sentence-transformer embeddings, persistent vector store, and Exa web search.

Culture ship names as Claude Code spinner verbs. Because why would you settle for 'Thinking...' when you could have 'Experiencing A Significant Gravitas Shortfall'?

Automated UI/UX design analysis — extracts W3C DTCG tokens, component inventory, and layout analysis from any URL.

Skill authoring toolkit — unified skill creation, TDD-adapted testing, and audit. Consolidated from interdev's create-agent-skills and writing-skills.

Slack integration for Claude Code — send messages, read channels, test integrations.

Agent and skill performance profiler and routing optimizer. Collects evidence about flux-drive agent accuracy and skill quality, proposes routing overrides and skill overlays for underperformers, and monitors canary periods. Companion plugin for Clavain.

Token efficiency benchmarking, session analytics, and API-equivalent cost analysis for agent workflows

Multi-agent synthesis engine — collects findings from parallel review/research agents, deduplicates, writes verdicts, produces compact summaries. Keeps agent output out of the host context.

Engineering quality disciplines — systematic debugging, test-driven development, and verification gates.

Cross-module integration gap tracer — traces data flows from shipped features to find unverified consumer edges.

Project hierarchy management — filesystem discovery, parent-child relationships, tagging, and layout orchestration

Agent trust scoring — reputation tracking, severity-weighted decay, and suppression candidates.

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. Surfaces drift on doc access via a PreToolUse hook with auto-fire on Certain confidence (cooldown + daily budget guarded). Companion plugin for Clavain.

Tool usage analytics for Claude Code. Tracks tool patterns via hooks, detects inefficiencies, and offers community comparison with anonymized data.

TUI automation and visual testing - Playwright for terminal applications. Spawn, interact with, and screenshot TUI apps.