Generic assistants lose context across conversations and notes. This skill uses local personas, notes, logs, and vector search to make Claude, Codex, or Claude Code respond with more personal context.

Complex requests often span planning, writing, review, and implementation. This skill selects relevant installed skills and presents a concise working set.

Cypress questions can produce incorrect tests when answers rely on memory or outdated examples. This skill finds official documentation and returns structured, source-grounded guidance.

Writing stable Cypress tests requires project context, framework knowledge, and careful command usage. This skill creates, updates, and repairs tests using established project conventions.

Technical writing often becomes wordy, vague, or hard to scan. This skill applies concise style rules to drafts, documentation, messages, and user-facing copy.

Outdated documentation slows reviews and creates wrong project assumptions. This skill audits docs against the codebase and preserves important context while making factual updates.

Complex architecture, debugging, and review questions often require evidence from code and documentation. The Oracle guides a Claude, Codex, or Claude Code agent through structured research and clear synthesis.

Engineering teams lose rationale when choices stay only in code review. The Archivist helps Claude, Codex, and Claude Code create ADRs, decision logs, and inline rationale.

Zustand stores often become inconsistent when actions, persistence, and updates follow different patterns. This skill guides Claude, Codex, and Claude Code toward one StoreBuilder approach.

Code changes can drift from plans, miss acceptance criteria, or introduce avoidable defects. This skill guides Claude, Codex, and Claude Code through structured diff review and merge-readiness checks.

Complex project plans become hard to navigate when details live in one long file. This skill creates linked Obsidian markdown wikis with focused tasks, research notes, and changelog records.

Vague issues slow implementation and lose investigation context. This skill turns findings, requirements, and acceptance criteria into complete GitHub tickets.

Inconsistent filenames make repositories harder to scan, review, and maintain. This skill audits naming rules, finds mismatches, and produces a prioritized cleanup report.

Coordinated agent work can miss issue requirements or merge incomplete PRs. Devflow provides gates for task extraction, review, approval, and issue-linked merges.

Manual design handoff from screenshots is slow and inconsistent. This skill extracts tokens, components, layout, and accessibility notes into structured design-spec files.

Complex questions often benefit from independent investigation paths. Delphi runs multiple identical oracle prompts and synthesizes their reports into one clear answer.

Teams often need specialized AI assistants, but subagent setup can be inconsistent and over-permissive. This skill provides templates, configuration guidance, and delegation patterns for reusable Claude Code subagents.

Large code changes fail when requirements, files, and validation steps are unclear. This skill guides Claude, Codex, and Claude Code to produce structured implementation plans before work starts.

Code changes can hide runtime failures, missing requirements, and weak verification plans. This skill guides Claude, Codex, or Claude Code through a structured review before merge.

Building a skill around an MCP server requires reliable discovery, testing, and documentation. This skill gives Claude, Codex, and Claude Code a structured workflow for MCP skill creation.