Ensure accessibility in UI components including semantic HTML, ARIA attributes, keyboard navigation, and WCAG 2.2 AA compliance.
日本語の概要は準備中です。原文の説明を表示しています。
Start Telegram monitoring. Auto-enables voice pipeline if ElevenLabs or OpenAI TTS keys are detected.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
Read .claude/context/runtime/channel-daemon.pid. If it exists and has content, verify the PID is alive using node -e "try { process.kill(PID, 0); console.log('ALIVE') } catch { console.log('DEAD') }". If ALIVE, tell the user the daemon is already running and show them status commands. Done — do NOT try to start again.
Spawn a developer agent:
subagent_type: developer
prompt: Run this command and report the output: node scripts/channels/telegram-ctl.cjs start
Read .env and check for ELEVENLABS_API_KEY or OPENAI_API_KEY. If either is set, tell the user:
"Voice pipeline auto-enabled — voice messages will be transcribed (Whisper) and responded to with audio (ElevenLabs/OpenAI TTS)."
If neither is set, just say:
"Text-only mode. Add ELEVENLABS_API_KEY or OPENAI_API_KEY to .env for voice message support."
Tell the user what's active:
/ in Telegram for the menunode -e with process.kill(pid, 0) or http.get for health checksまだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Ensure accessibility in UI components including semantic HTML, ARIA attributes, keyboard navigation, and WCAG 2.2 AA compliance.
日本語の概要は準備中です。原文の説明を表示しています。
Use when you want to improve response quality through meta-cognitive reasoning. Applies 15+ reasoning methods to reconsider and refine initial outputs.
日本語の概要は準備中です。原文の説明を表示しています。
N-round opposing-stance debates for trade-off analysis. Assigns pro/con roles to agents, runs structured debate rounds with quality scoring, and produces a moderator synthesis with confidence-rated recommendation. Generalizable to architecture, technology, security, and design decisions.
日本語の概要は準備中です。原文の説明を表示しています。
Force adversarial code review stance that eliminates confirmation bias — reviewer must find issues or re-analyze
日本語の概要は準備中です。原文の説明を表示しています。
Creates specialized AI agents on-demand when no existing agent matches a request. Use when the Router cannot find a suitable agent for a task. Enables self-evolution by generating persistent agents.
日本語の概要は準備中です。原文の説明を表示しています。
LLM-as-judge evaluation framework with 5-dimension rubric (accuracy, groundedness, coherence, completeness, helpfulness) for scoring AI-generated content quality with weighted composite scores and evidence citations
日本語の概要は準備中です。原文の説明を表示しています。