harness-session

Plugin harness-session in modakbul-gongbang/harness. Commands read from the repo's manifest — 1 month ago.

Skills (7)

• agent-orchestrate

  Analyze the user's task and propose the optimal orchestration pattern, then execute it. 4 patterns: Sequential Pipeline, Parallel Subagent, Team Mode, Ralph Loop. Situation-aware pattern selection with user confirmation before execution. Use when: "/agent-orchestrate", "agent-orchestrate", "오케스트레이션"

• check-harness

  Audit a Claude Code project's development readiness: runtime and LSP, context, permissions, hooks, tests/CI, skills and independent verification. Distinguish configured, observed working, missing and unverified capabilities. Use for /check-harness, 하네스 점검, LSP 설치 확인, or 개발 환경 종합 점검.

• deep-interview

  /deep-interview", "deep interview", "interview me", "clarify requirements", "요구사항 정리", "인터뷰", "딥 인터뷰", "뭘 만들어야 할지 모르겠어", "요구사항이 불명확", "아이디어 구체화" - Read - Grep - Glob - Agent - Write - WebSearch - AskUserQuestion Must contain all 3 stages: INITIATE, INTERVIEW, SYNTHESIZE. Must launch interviewer agen

• doc-drift

  Use this skill when the user wants to audit the memory and documents Claude Code loads into context — CLAUDE.md (user global + project + nested), MEMORY.md, @imports, .claude/skills, .claude/agents, .claude/commands, installed plugins — and detect three kinds of issues: outdated claims, mutually con

• qa

  Systematically QA test any application — web apps, native macOS apps, Electron apps, CLI tools, interactive REPLs, or anything on screen. Three modes: browser (agent-browser/Playwright, fast, DOM-level), computer (MCP computer-use, screenshot + pixel clicks, any app), and cli (tmux, send-keys + capt

• scaffold

  Design and optionally build a new project's architecture and AI work environment. Select a stack and dependencies, plan a working reference flow, and configure project instructions, verification, and useful skills/hooks. Use for "/scaffold", "new project setup", "프로젝트 세팅", or "초기 구조". Existing proje

• specify

  Turn a goal into an implementation plan (spec.md). Simplified layer chain: L0:Goal → L1:Context → L2:Decisions → L3:Requirements → L4:Tasks. Evidence-based clarity scoring at L2. User approves at L2, L3, L4. Output is a single spec.md file written with the Write tool. Use when: "/specify", "specify"

Subagents (4)

• context-quality-reviewer

  Read-only context and permission audit; separates configuration from runtime evidence.

• project-automation-auditor

  Read-only development tooling and verification audit; separates configuration from runtime evidence.

• session-pattern-analyzer

  Analyzes Claude Code session JSONL files to extract execution patterns: plan-ratio, delegation, parallel usage, handoff, repeated n-grams, tool frequency. Uses jq for efficient extraction of tool_use metadata only — never reads prompt text. Scales via split+parallel when many sessions exist. Returns

• skill-portfolio-analyzer

  Analyzes installed skills/plugins/MCP servers against ~/.claude.json to detect dead, ghost, duplicate, and collision-prone entries AND report the current accessible state (enabled plugins, runtime skills, connected MCP servers). Returns a structured JSON report. Read-only. Use from /check-harness. -