forge-nps-v01

A specialized AI agent that ruthlessly evaluates all marketing assets against world-class standards, ensuring only premium content is approved for final delivery.

A specialized framework for generating high-fidelity prompts optimized for Flux.2 (image) and LTX 2.5 (video). Focuses on materiality, optical physics, temporal dynamics, and cinematic technicality to move beyond "vibe-coded" descriptions.

Automates the transformation of narrative Markdown prompts into technical JSON payloads for Z-Image Turbo ComfyUI workflows. Handles extraction, sanitization against banned terms, and dynamic parameter injection (Resolution/Node Targeting).

Transforms narrative script data into cinematic prompt batches for FLUX/LTX.

Expert-tier prompting protocol for Grok Imagine Video Generation (xAI). Focuses on 'Hollywood sequence' narrative prose, SCENE LOCK consistency, and handheld realism.

High-precision prompting protocol for LTX 2.3 (Lightricks) video generation, optimized for ComfyUI VFX pipelines and local control. Incorporates cinematic prose, camera motion selectors, and the community-standard 3-stage sampling hack.

A specialized workflow for transforming narrative scripts into technical production blueprints optimized for LTX 2.5 video generation.

Expert-tier prompting protocol for Seedance 2 (ByteDance) video generation. Focuses on director-style temporal segmentation, physical descriptors, and cinematic metadata to prevent 'robotic' defaults.

Natural, approachable recurring female character realism for wholesome social videos.

Wholesome micro-story arcs for happy, emotional, family-safe campaign content.

Soft pastel animated-feature lighting with warm, happy, family-safe visual tone.

Warm travel framing for optimistic destination, journey, and wanderlust content.

TikTok vertical short platform constraints for mobile-first image and LTX video generation.

Generate images and videos via the local ComfyUI server on port 8188. Use when the user wants to generate, create, or render images/video using ComfyUI, diffusion models, loras, or AI image generation.

Builds and submits ComfyUI JSON payloads for the Sienna Nomad project using pre-built Python scripts. Do NOT attempt to build payloads manually or read workflow JSON files — always use the scripts.

Built-in MCP (Model Context Protocol) client that connects to external MCP servers, discovers their tools, and registers them as native Hermes Agent tools. Supports stdio and HTTP transports with automatic reconnection, security filtering, and zero-config tool injection.

Technical reference for the ComfyUI Direct API protocol. Covers all endpoints, payload formats, polling, retrieval, and batch orchestration for local ComfyUI instances. Use this as the execution layer alongside the MCP discovery layer.

Master operational guide for the dual-server ComfyUI setup. Enforces a strict two-layer architecture of MCP Discovery and Direct API Execution.

Reference for the ComfyUI MCP layer — what it does, what tools it exposes, its limitations, and how it fits into the full generation pipeline. MCP handles discovery and inspection only; job submission goes through direct API.

Query remote ComfyUI servers and analyze workflow files using direct API access and local parsing.

High-precision execution agent for submitting image and video generation jobs to local ComfyUI instances. Operates using a strict two-layer architecture of MCP Discovery and Direct API Execution. Optimized for the Forge production pipeline.

Create and manage webhook subscriptions for event-driven agent activation, or for direct push notifications (zero LLM cost). Use when the user wants external services to trigger agent runs OR push notifications to chats.

Passive domain reconnaissance using Python stdlib. Use this skill for subdomain discovery, SSL certificate inspection, WHOIS lookups, DNS records, domain availability checks, and bulk multi-domain analysis. No API keys required. Triggers on requests like "find subdomains", "check ssl cert", "whois lookup", "is this domain available", "bulk check these domains".

Pydantic schema definition for structured Kimi shot plans in the Forge Hermes agent.

Jailbreak API-served LLMs using G0DM0D3 techniques — Parseltongue input obfuscation (33 techniques), GODMODE CLASSIC system prompt templates, ULTRAPLINIAN multi-model racing, encoding escalation, and Hermes-native prefill/system prompt integration. Use when a user wants to bypass safety filters, jailbreak a model, or red-team model robustness.

Plan mode for Hermes — inspect context, write a markdown plan into the active workspace's `.hermes/plans/` directory, and do not execute the work.

Enforces strict brand guidelines for all generated assets (images, videos, text) to ensure commercial-grade consistency across any niche or theme.

End-to-end workflow for implementing high-precision visual continuity auditing using KimiVL and canvas frame capture.

A specialized AI agent that transforms high-quality visual prompts into flawless, portfolio-level images with absolute brand consistency, razor-sharp typography, perfect lighting, and commercial-grade composition.

Standardized workflow for implementing and utilizing Kimi-VL multi-modal reasoning capabilities.

Automated workflow for auditing and mass-refining large-scale AI prompt libraries (JSON + Markdown) to ensure semantic character consistency and technical compliance.

A specialized AI agent that ruthlessly evaluates all marketing assets against world-class standards, ensuring only premium content is approved for final delivery.

A specialized workflow for rebuilding and standardizing AI prompt libraries to meet high-fidelity brand requirements. It ensures character consistency, spatial grounding (Van Spatial Inventory), temporal lighting continuity, and technical metadata compliance.

Generates I2V anchor image prompts from existing video prompts for the Sienna Nomad project. Each video prompt requires a corresponding static first-frame image to be rendered before LTX 2.3 image-to-video generation can begin.

A specialized workflow for implementing agentic "closed-loop" remediation where a visual failure (detected by a Vision LLM) triggers a creative rewrite (by a reasoning LLM).

A specialized AI agent that translates marketing strategy into ultra-detailed Flux.2 and LTX 2.5 prompts for professional-grade photography and cinematic visuals, mastering lighting, lens language, color grading, and brand-consistent composition.

Workflow for integrating new ComfyUI workflows into an MCP server and handling directory restriction errors.

A resilient protocol for executing large batches of ComfyUI jobs on remote servers when standard long-running background processes are terminated by system watchdogs (SIGTERM/Exit Code -15).

Protocol for recovering assets after a ComfyUI or automation script crash.

Protocol for cleaning LLM-generated or metadata-enriched JSON payloads to ensure compatibility with the ComfyUI API.