Full-cycle delivery orchestrator for AI coding agents: a mandatory intake grill that front-loads every decision, then ten gated stages — docs, brainstorm, decompose, spec, plan, build, tests, deploy, post-deploy, acceptance — with a loop guard that breaks churn. Loads in DeepSeek Harness (dsh).
Eight agent skills, one command, every agent. Umbrella installer for the ssheleg family — super-ux, task-pipeline, agent-sync, make-skill, sheleg-design, seo-aeo-audit, sheleg-dev, agent-stack — across Claude Code, Cursor, Codex and 70+ more agents. Loads in DeepSeek Harness (dsh).
Evidence-first website audit for search AND answer engines. Ten tracks from crawl access to AI citation mechanics; every finding carries an observation, every recommendation an evidence tier. Output is a prioritized change plan plus a link-building brief with keyword CSV — not a score. Loads in DeepSeek Harness (dsh).
Scenario-driven UI development for AI coding agents: a versioned design chain in docs/ux/ — personas and jobs → user flows → a screens-and-states map with Figma frames → traced scenarios → evidence-backed audits → fix plans. One /ux entry point, a linter that fails when the docs drift from the code. Loads in DeepSeek Harness (dsh).
Several coding agents, one repository, no collisions — leases with a TTL, race-free id reservation, a run journal and a generated board over a pluggable knowledge cloud. Ships in the sshlg-skills bundle. Loads in DeepSeek Harness (dsh).
A skill that builds skills: create, audit and ship Agent Skills and Claude Code plugins — conformance to the open standard, a validator that can fail, every distribution channel, and the MCP/A2A boundary. Loads in DeepSeek Harness (dsh).
The taste layer for AI coding agents: cinematic scroll-driven landing pages (one scroll clock, motion that degrades to calm, WebGL particle formations), product-UI style packs each shipping a ready token layer, and the Figma border — tokens as variables, design to code without hand-copied values. Loads in DeepSeek Harness (dsh).
Money in, tracking, sign-in, speed — the integration layer a product reaches once it has users. Seven agent skills: Stripe subscription billing reconciled into your own database, crypto payments, GA4/Ads/Meta under Consent Mode v2, Google sign-in, Core Web Vitals, Sentry error tracking. Loads in DeepSeek Harness (dsh).
Production patterns for agent systems, in four agent skills: orchestrator loops and memory layers, the harness (system prompts, tool descriptions, technique choice), evals that judge trajectories rather than answers, MCP and A2A interop — plus the wallet under reselling LLM access. Loads in DeepSeek Harness (dsh).
Telegram development split by the API each surface speaks: the official Bot API for bots, MTProto/Telethon for user accounts, and Mini Apps for the web layer whose whole security model is one signed blob. Three agent skills for Claude Code, Cursor and 70+ agents. Loads in DeepSeek Harness (dsh).
