terraphim

Log exploration and analysis using Quickwit search engine. Incident investigation, error pattern analysis, and observability workflows. Three index discovery modes for different performance and convenience trade-offs.

Search and analyze AI coding assistant session history using Terraphim. Find past conversations, discover patterns, and learn from previous work. Supports Claude Code, Cursor, Aider, and other AI coding assistants.

Plan and (when feasible) implement or execute user acceptance tests (UAT) / end-to-end acceptance scenarios. Converts requirements or user stories into acceptance criteria, test cases, test data, and a sign-off checklist; suggests automation (Playwright/Cypress for web, golden/snapshot tests for CLIs/APIs). Use when validating user-visible behavior for a release, or mapping requirements to acceptance coverage.

System architecture design for Rust/WebAssembly projects. Creates ADRs, designs APIs, plans module structures, and documents architectural decisions. Never writes implementation code - focuses purely on design and documentation.

Thorough code review for Rust/WebAssembly projects. Identifies bugs, security issues, performance problems, and maintainability concerns. Provides actionable feedback with specific suggestions.

Open source community building and engagement. Welcoming contributors, managing discussions, writing release notes, and fostering a healthy project ecosystem.

Systematic debugging for Rust applications. Root cause analysis, logging strategies, profiling, and issue reproduction. All debug changes removed before final report.

DevOps automation for Rust projects. CI/CD pipelines, container builds, deployment automation, and infrastructure as code. Optimized for GitHub Actions and Cloudflare deployment.

Secure secret management using 1Password CLI. Detect plaintext secrets in files and codebases, convert environment files to 1Password templates, inject secrets securely using op inject, and audit codebases for security compliance.

Comprehensive Caddy web server management across multiple environments. Handles multi-server operations, zero-downtime deployments, secret management with 1Password, custom builds with plugins, and systemd service management.

Generates Rust code for GPUI desktop UI components following Zed editor patterns. Use when building desktop applications with gpui crate, creating themed UI components, implementing autocomplete/completions, building command palettes, or working with the gpui-component library. Covers RenderOnce components, Entity state management, theming with ActiveTheme, and Zed-style UI patterns.

MD-Book documentation generator skill. Use when working with MD-Book, a modern mdbook replacement for generating HTML documentation from Markdown files. Supports syntax highlighting, live reload, Pagefind search, and Tera templates.

Phase 2 of disciplined development. Creates implementation plans based on approved research. Specifies file changes, function signatures, test strategy, and step sequence. Requires human approval before implementation.

Phase 3 of disciplined development. Executes approved implementation plans step by step. Each step includes tests, follows the design exactly, and produces reviewable commits.

Phase 1.5/2.5 of disciplined development. Evaluates document quality using the KLS (Krogstie-Lindland-Sindre) 6-dimension framework. Produces structured ratings, identifies gaps, suggests specific revisions, and blocks phase transitions when quality is below threshold. Applies to Phase 1 research docs and Phase 2 design docs.

Phase 1 of disciplined development. Deep problem understanding before design. Maps existing systems, identifies constraints, surfaces risks, and produces a research document for human approval before proceeding.

Query and analyse agent evolution data from the ADF orchestrator. Use when the user asks about "agent evolution", "agent memory", "evolution snapshot", "agent learning history", or "agent performance over time". Triggers: "evolution", "agent memory", "agent learning", "snapshot".

Recursive Language Model (RLM) orchestration for Terraphim AI. Provides isolated code execution in Firecracker VMs, session management, budget tracking, and knowledge graph validation. Use when: (1) Executing LLM-generated code in sandboxed environments, (2) Running untrusted code with resource limits, (3) Managing long-running code execution sessions with snapshots, (4) Building AI agents that write and execute code. Triggers: "execute code", "run python", "sandbox", "code isolation", "RLM", "firecracker", "session management", "budget tracking".

Automatically replace package manager commands (npm, yarn, pnpm) with bun using Terraphim's knowledge graph system. Use this when the user mentions package managers, installation commands, or build scripts that use npm/yarn/pnpm.

PageRank-powered task management using Gitea, tea CLI, and gitea-robot. Replaces beads/br and mcp_agent_mail with Gitea-native tooling. Use when agents need to pick work, prioritize issues, manage dependencies, coordinate multi-agent work, or run agile workflows against git.terraphim.cloud. Triggers: 'what should I work on', 'pick next task', 'triage issues', 'show ready tasks', 'prioritize backlog', 'claim issue', 'close issue', 'add dependency', 'show dependency graph', 'coordinate work', 'start session', 'run workflow'. Keywords: gitea, pagerank, triage, ready, robot, tea, issue, priority, dependency, backlog, sprint.

Design system for Terraphim.ai enforcing radical essentialism. Combines neumorphic depth, purposeful motion, and expanding grid layouts. Use when creating UI components, pages, or applications. Triggers: (1) Building web interfaces, (2) Creating design systems, (3) Implementing neumorphic UI, (4) Adding purposeful animations, (5) Reviewing designs for brand alignment. Provides design tokens, component patterns, layout utilities, and quality checklists. Modern CSS with oklch colors, CSS custom properties, container queries, and View Transitions API.