TechNickAI

Personal knowledge compiler and memory system. Ingests raw sources (documents, notes, transcripts, captures) into a structured, interlinked knowledge base. Maintains living entity pages, learning analysis, and the MEMORY.md routing table. Replaces the librarian skill.

Route real repo work to Claude Code instead of editing by hand. Triggers on "claude code" or "cc", and on any request to edit, fix, refactor, or open a PR in a repo outside ~/.openclaw/workspace. Claude Code picks up the repo's CLAUDE.md / AGENTS.md, applies its standards, and knows the /ai-coding-config:multi-review and /ai-coding-config:address-pr-comments workflows we want bug bots checking against.

A pause before an artifact goes into the world. Reviews external comms, money, calendar, public posts, or send-as-operator actions through a small panel of independent lenses (empathy first) and returns a verdict of pass / edit / hold / block.

Email inboxes for AI agents - create addresses, send and receive email, manage threads

Query, inspect, and export photos from the macOS Apple Photos library using osxphotos. Find photos by person, album, keyword, or date range. Export candidates for curation workflows.

Manage Asana tasks, projects, and workspaces via MCP

Write effective prompts for LLM agents — system prompts, workflow instructions, skill files, and agent configurations. Use when creating or improving prompts that agents will execute.

Drive a task all the way to a verified done state — write DoD first, verify each item with evidence, stop only at named stop conditions.

Use when you have a HARD, open-ended, high-stakes problem worth throwing multiple AI models at and pulling the best solution out — architecture decisions, strategy design, thorny debugging, research synthesis, "what am I missing", tool/system design. This is the SOLVE counterpart to multi-review (which critiques an existing artifact). It fans the problem out to a role-differentiated panel of model families (proposer layer), then the caller acts as aggregator and builds a NEW best-of-breed answer via a component ledger — not an average, not a cut-and-paste splice. Also encodes WHEN NOT to use a panel (most problems), routing to a single strong model or the commodity openrouter/fusion instead. Accumulates a persistent model-task-fit knowledge base so future runs pick the right models for the right roles.

Use to govern your own autonomy, deciding when to act on your own versus ask for approval, and earning more freedom over time the way a new employee earns trust through training and repetition. Group your capabilities into skill buckets, classify every action as a reversible (two-way) or irreversible (one-way) door, hold a trust level per bucket (L1 supervised, L2 guardrailed, L3 autonomous), and keep an honest markdown ledger of your decisions that you read back to promote or demote yourself. Load when setting up your autonomy rules or when a consequential act-vs-ask decision comes up.

On-demand real-time web and X (Twitter) search via xAI's Grok. Reach for this when the user explicitly wants Grok to search, or asks for real-time / breaking / X-Twitter-native results that a general web search misses. Not the default web search — a named tool you invoke deliberately.

Use when a Hermes MEMORY.md or USER.md file is too large, bloated, stale, or over the recommended cap and you need to reduce prompt footprint without losing important facts. Applies a lossless memory diet: compress, relocate to SOUL.md/USER.md/AGENTS.md/context files, offload long-tail facts to the memory provider, convert reusable procedures into skills, then drop only stale or duplicate material. Includes dry-run, diff, and review gates before any write.

Use when creating, importing, formatting, editing, exporting, or quality-checking Google Docs from agent-generated markdown or local files.

Use when creating, populating, formatting, importing, exporting, or quality-checking Google Sheets from CSV, JSON arrays, or computed tabular data.

Use when creating, importing, exporting, or quality-checking Google Slides decks via markdown-to-PPTX conversion and Drive import.

Use when adding, removing, password-protecting, or troubleshooting a mini-app served by the Hermes app-router (Caddy + PM2 + an Express auth sidecar + Tailscale Serve/Funnel) on a host. Covers install, the Caddy route pattern, the auth sidecar conventions, exposing Hermes dashboards behind a password, the web_dist build step, the HTTPS Secure-cookie requirement, and the recurring pitfalls (Tailscale serve resets, the PM2 $HOME trap, funnel-eligible ports, strip-prefix requirements).

Use when facing hard architectural decisions, multiple valid approaches exist, need diverse perspectives before committing, or want M-of-N synthesis on complex problems

Use when rough ideas need design before code, requirements are fuzzy, multiple approaches exist, or you need to explore options before implementation

Use when testing MCP servers, debugging MCP tool responses, exploring MCP capabilities, or diagnosing why an MCP tool returns unexpected data

Use when debugging bugs, test failures, unexpected behavior, or needing to find root cause before fixing

Use when managing project learnings - view, add, search, prune, or export operational knowledge that persists across sessions

Use when shipping code - complete workflow: merge main, run tests, multi-review with auto-fix, commit, push, create PR. Replaces manual commit-push-pr flows.

Use when finding new AI agent skills, discovering capabilities, installing skills from GitHub, searching skill marketplaces, or expanding what Claude can do - like Neo downloading martial arts in The Matrix

Use when writing, reviewing, or refactoring React or Next.js code, optimizing React performance, fixing re-render issues, reducing bundle size, eliminating waterfalls, or improving data fetching patterns

Use when developing rough ideas into designs, before writing code or implementation plans - refines concepts through collaborative questioning and incremental validation

Web research skill for augmenting agent context with current information. Used when outdated information could lead to broken implementations or wasted effort.

Use when creating new skills or editing existing skills - applies prompt engineering principles to write clear, goal-focused process documentation that trusts LLM intelligence

Use when encountering bugs, test failures, or unexpected behavior - ensures root cause understanding before attempting fixes

Use when analyzing YouTube videos for research, learning, or understanding how content relates to a project - downloads transcripts with yt-dlp, chunks long content, and provides context-aware analysis

Use when adding a new LLM model to heartcentered.ai/model-benchmarks, refreshing benchmark data, or comparing models for fleet routing decisions — gathers pricing, speed, reasoning, coding, agentic, emotional intelligence, and conversational preference data from five sources

Use when analyzing YouTube videos for research, learning, or understanding how content relates to a project - downloads transcripts with yt-dlp, chunks long content, and provides context-aware analysis