iclaude

Use when improving ONE numeric metric under a controlled research loop with a fixed eval and kept/reverted experiments. Not for a feature (use loop-delivery) or a failing test (use loop-repair).

Use when delivering ONE bounded change — a feature, refactor, or chore — as a controlled, audited loop in any repo. Not for a failing test (use loop-repair) or a numeric metric (use loop-autoresearch).

Use when a specific test, CI job, or regression is failing and must be fixed under a reproduce-first controlled loop with proven regression coverage. Not for delivering a change (use loop-delivery) or metrics (use loop-autoresearch).

Use when a loen loop stage — plan, act, check, or result — must be validated and gated before the next one. Mode-aware for delivery/repair/research; the execution-loop analog of check-chain.

Use when you need a cross-run dashboard over all docs/loen/ runs, or --triage to turn failing runs into proposed next actions (proposals only; never launches loops or edits runs).

Use when an active, human-approved loen run should keep going multi-turn on its own — wraps it in Claude's native /goal from loop.yaml. Optional; never bootstraps a run or submits /goal itself.

Use BEFORE superpowers:brainstorming for any non-trivial work (new module, CLI flag, API change, arch decision). Triggers on "/fix-intent", "/idd", "capture intent", "intent doc". If brainstorming would start on non-trivial work without an approved intent doc in docs/superpowers/intents/, run IDD first. Brainstorm proceeds only when intent Status: approved.

Use to validate the IDD→SDD chain (intent → spec → plan → result). Triggers on "/check-chain", "check chain", "validate intent/spec/plan/result", and is the remediation the chain-gate hook points to. Runs the whole chain (sequential gate) with no argument, or a single stage with "/check-chain <stage>".

Верификация и переписывание СУЩЕСТВУЮЩИХ инструкционных файлов (CLAUDE.md, AGENT.md, SKILL.md) против 7 правил форматирования. Использовать когда пользователь просит "проверить/исправить/проаудитить промт", "агент игнорирует правила", "отрефакторить инструкции", перед коммитом изменений в инструкционные файлы. НЕ для создания новых агентов — использовать agent-builder.

Detect project language, framework, package manager, lint/test commands and locate CLAUDE.md / PRD docs at task start (Phase 0). Also detects the iwiki MCP domain for this project, surfacing its summary as project context. Use when starting any task, switching project, or before running syntax/test checks. NOT for deep semantic doc search (wiki_search) — this skill only detects availability + a quick summary.

Answer a question from the project's docs/wiki/ using semantic embedding search. Use when the user asks "where/how/why" about the project, or when CLAUDE.md says to consult the wiki before starting work.

Generate or update a wiki page in docs/wiki/ from a source file or folder, then refresh the embedding index. Use when asked to document code, add to the wiki, or after changing functionality. GUARDED: always show a diff and log.

Bootstrap a project's docs/wiki/ from scratch — scan the source tree, generate one wiki page per major area, build the embedding index, and lint. Use for a brand-new wiki (empty or missing docs/wiki/). GUARDED: shows the area plan and a diff before writing; sending source to the embedding API is consented egress.

Report documentation health over docs/wiki/: broken [[refs]], orphan pages, stale pages, and gaps (source with no wiki page). Report-only — makes no edits.

Use when the user asks for a standalone or self-contained HTML report, an offline .html file, a human-readable report that opens by double-click in a browser, expandable/interactive HTML tables, or block / transition / C4 / architectural diagrams rendered in HTML (NOT Mermaid), or a dark/light themed report. Triggers on phrases like 'сделай html-отчёт', 'standalone html report', 'expandable html table', 'C4 diagram as html', 'offline report I can open in browser'. Produces ONE self-contained .html in docs/reports/ with zero external dependencies. NOT for Mermaid diagrams (use mermaid-obsidian) or Mermaid embedded in PRD/architecture docs (use prd-generator / architecture-documentation).

Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.

Standardized git workflow with Conventional Commits. Use when creating a feature branch, staging commits, opening a PR, or when user says "commit", "create branch", "open PR", "fix commit message". Enforces commit prefix (feat/fix/docs/...), branch naming, PR template.

Use this skill whenever you need to create, generate, write, or fix Mermaid diagrams for viewing in Obsidian. Triggers when the user asks to "draw a diagram", "create a flowchart", "make a sequence diagram", "visualize architecture", "show relationships as a diagram", "create a mindmap in Obsidian", mentions Mermaid syntax, asks to fix a broken diagram, or wants to add any kind of chart/graph to their Obsidian notes. Covers all Mermaid diagram types: flowchart, sequence, class, ER, state, gantt, pie, gitGraph, mindmap, timeline. Always use this skill when user asks to draw/fix a standalone diagram in Obsidian. NOT for diagrams embedded inside PRD documents (use prd-generator) or architecture YAML (use architecture-documentation) — those skills generate Mermaid internally.

This skill should be used when the user asks to "analyze session", "summarize session", "compact session", "what happened in session", "/compact-session", or provides a path to a .jsonl session file and wants a summary of what was done.

Централизованный API для конвертации JSON ↔ TOON и расчёта token savings