verivus-oss

async-job-orchestration

Manage long-running async LLM jobs. Use for tasks >2min, parallel jobs, or non-blocking execution. Covers cache-aware `promptParts`, `cache-state://` MCP resources, and the Claude `cache_ttl_expiring_soon` warning.

multi-llm-review

Parallel code reviews across Claude, Codex, Gemini, Grok, and Mistral. Use for quality analysis, bug finding, or security audit.

secure-orchestration

Security-conscious LLM orchestration with approval gates across Claude, Codex, Gemini, Grok, and Mistral. Use for high-risk operations, permissions, auditing.

session-workflow

Manage conversation sessions across Claude, Codex, Gemini, Grok, and Mistral. Use for multi-turn conversations, session switching, workspace management. Covers the `session_get.cacheState` projection, cache-aware `promptParts`, and `cache-state://` MCP resources.

implement-review-fix

Run implement-review-fix cycle using multiple LLMs (Claude, Codex, Gemini, Grok, Mistral). Use for features, bugs, or refactoring with multi-LLM collaboration. Mistral defaults to `--agent auto-approve`; current Vibe defaults session logging on, and doctor flags explicit `[session_logging] enabled = false` before session-continuity use.

provider-claude

Track and maintain the upstream Claude Code CLI contract. Use when Anthropic ships a Claude Code release, when a `claude` flag/output/permission behaviour changes, or when an upstream scan flags drift. Process guidance only — `src/upstream-contracts.ts` is the mechanical source of truth.

provider-codex

Track and maintain the upstream OpenAI Codex CLI contract. Use when OpenAI ships a Codex release, when a `codex exec` flag/sandbox/approval/resume behaviour changes, or when an upstream scan flags drift. Process guidance only — `src/upstream-contracts.ts` is the mechanical source of truth.

provider-gemini

Track and maintain the upstream Google Gemini CLI contract. Use when Google ships a Gemini CLI release, when a `gemini` flag/approval-mode/output-format/resume behaviour changes, or when an upstream scan flags drift. Process guidance only — `src/upstream-contracts.ts` is the mechanical source of truth.

provider-grok

Track and maintain the upstream xAI Grok CLI contract. Use when xAI ships a Grok CLI release, when a `grok` flag/permission-mode/sandbox/output-format/resume behaviour changes, or when an upstream scan flags drift. Process guidance only — `src/upstream-contracts.ts` is the mechanical source of truth.

provider-mistral

Track and maintain the upstream Mistral Vibe CLI contract. Use when Mistral ships a Vibe release, when a `vibe` flag/agent-mode/output-format/session-logging/env-model behaviour changes, or when an upstream scan flags drift. Process guidance only — `src/upstream-contracts.ts` is the mechanical source of truth.

public-demo-session

Prepare and drive clean public llm-cli-gateway demo sessions for recordings, README demos, screenshots, or transcripts. Use when the user needs a redacted path, compact Codex output, direct provider calls, or a demo-safe Codex environment.

agent-codex-gate

Pattern for spawning subagents that must get Codex approval before their work is accepted. Use when dispatching parallel agents that each need independent Codex review gates.

codex-review-gate

Submit work to Codex for review and iterate until unconditional approval. Use after completing implementation tasks, before merging, or when a development process requires Codex sign-off.

design-review-cycle

Structured design document review via LLM gateway — submit plans, specs, or designs for peer review from Codex, Gemini, Grok, or Mistral, iterate on feedback, track review rounds. Use before implementing complex features. Mistral Vibe defaults to `--agent auto-approve`.

model-routing

Choose the right LLM and model for each task based on proven patterns. Use when deciding whether to delegate to Claude, Codex, Gemini, Grok, or Mistral, or when selecting model variants. Mistral Vibe selects model via `VIBE_ACTIVE_MODEL` env var (no `--model` flag).

multi-llm-consensus

Run a task through multiple LLMs (Claude, Codex, Gemini, Grok, Mistral) independently and require agreement before proceeding. Use for high-stakes generation, conflict resolution, or final quality gates requiring unanimous approval.

multi-llm-orchestration

Guide for orchestrating multiple LLMs via the llm-gateway — use when delegating tasks to Codex, Gemini, Grok, or Mistral, running parallel reviews, or managing cross-LLM workflows. Covers cache-aware `promptParts` dispatch and the `cache-state://` MCP resources.

red-team-assessment

Get an adversarial red team security assessment from any LLM (Claude, Codex, Gemini, Grok, or Mistral) with gateway-managed approvals and optional sqry, exa, and ref_tools MCP access. Use when you need adversarial security analysis of code, architecture, or configurations.