eta-mu-pi

Never run Docker containers as root. Enforce USER directive in Dockerfiles and user: in compose files. Database images are the only exception.

Handle packages migrating between pnpm workspaces. Clean stale root copies, update workspace configs, fix broken workspace:* references.

Recover lost source files from pi session history when files were deleted but never committed to git.

Improve the Hormuz clock model through new signal classes, schema changes, uncertainty modeling, and better rendering or branch logic

Revise fork-tax protocols so they assume concurrent agents by default in shared workspaces, forbid destructive repo-wide cleanup of unrelated dirt, and require path-scoped staging plus explicit blocker recording.

Persist the full working state into git (commit + tag + push + manifest artifacts) as a deterministic handoff snapshot. Use when the user requests Π / fork tax / full dump.

Standardize human login, service identity, and inter-service auth around AT Protocol DIDs and Bluesky-backed one-time sign-in, while removing host-local static secrets and bespoke per-service auth flows.

Safely source website credentials from the local ignored passwords.csv export for browser automation without printing or committing secrets.

Set up PR-based branch promotion workflows with tiered CI gates, staging/main deploy hooks, and explicit GitHub branch-protection follow-through.

SSH into Promethean hosts, inventory active Docker/Podman/systemd/Proxmox runtimes, map public subdomains to live services, and write both JSON records and a markdown report.

Choose staging and production Promethean host slots, subdomains, runtime paths, and compose-project names from the allowed base-host pool.

Create or update *.promethean.rest DNS A records through Cloudflare by copying current allowed base-host IPs while preserving unrelated zone records.

Interpret "Deploy X" as bootstrapping or repairing a full local -> PR -> staging -> PR -> prod Promethean deployment flow with DNS, GitHub automation, and live validation.

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

Skill: Create PM2 Configs

Skill: Create PM2 Ecosystem

Translate turn difficulty into explicit p-scores for efficiency, friction, and skill-candidate potential.

Skill: Giga Workflow

Skill: GitHub Integration

Normalize and curate Kanban/spec task files: ensure frontmatter, normalize tokens, and keep process vocabulary consistent.

Canonical Kanban workflow FSM contract (states, transitions, gates, invariants, normalization). Use to validate or explain status transitions.

Run a local Docker registry (localhost:5000), publish reusable CUDA-capable ML base images, and stand up local ML services when container/runtime issues appear.

Skill: LSP Server Integration

Skill: MCP Server Integration

Skill: Nx Integration

Skill: OpenCode Agent Authoring

Skill: OpenCode Agents & Skills

Skill: OpenCode Command Authoring

Skill: OpenCode Configs & Rules

Skill: OpenCode Model Variant Management

Skill: OpenCode Models & Providers

Skill: OpenCode Plugin Authoring

Skill: OpenCode Plugins

Skill: OpenCode SDK Workflows

Skill: OpenCode Session Search

Create OpenCode skills that validate and load correctly.

Skill: OpenCode Tool Authoring

Skill: OpenCode Tools, MCP, and ACP

Skill: PM2 Process Management

Start and stop a server with pm2 using `pm2 start ./my-server --name <name> --no-autorestart`, plus common supporting commands for status, logs, restart, and cleanup.