#١٣codejunkie99/agentic-harness١ skillsgreetGreeting workflow for the native hello-world agent١٢ مايو ٢٠٢٦
#١٤codejunkie99/familiar-second-brain١ skillsfamiliar-second-brainAlways use when working in Kimi Work desktop sessions that should read from or save to the user's Familiar second brain. Trigger for memory, notes, projects, prior work, research, summaries, durable outputs, "save this", "remember this", "from my brain",…١٤ يونيو ٢٠٢٦
#١٥codejunkie99/tokensavings-pro١ skillsdualmodeRun the Luna execution + Sol reasoning dual-model loop. Use when the user invokes $dualmode or /dualmode, asks to set up Luna + Sol handoffs, wants to split reasoning from execution to stretch Codex quota, or asks for the task packet loop. Resident in Codex…٣١ أغسطس ٢٠٢٦
#١٦codejunkie99/union-crew١ skillsunion-crewCoordinate work with Union Alpha as the orchestrator and DeepSeek V4.1 Flash subagents. Use when the user invokes /union-crew or $union-crew, or requests this specific model team.١٧ سبتمبر ٢٠٢٦
#١٧codejunkie99/voice-agent-builder١ skillsvoice-agent-builderBuild a production voice agent step by step. Use when the user wants to build, ship, or productionize a real-time voice AI agent — chained pipeline, speech-to-speech, or hybrid. Covers the full stack: streaming STT, LLM, RAG, TTS, function calling, safety,…١٨ مايو ٢٠٢٦
#١٨codejunkie99/quota-flow١ skillsquota-flowImplement features, fix bugs, create skills, and verify deliverables with Astra orchestration, Sol implementation, and narrowly scoped DeepSeek V4.1 Flash and Luna subagents while minimizing avoidable usage.١٢ سبتمبر ٢٠٢٦
#١٩codejunkie99/codex-model-routing١ skillsmodel-routingChoose a coding model and suggest effort using a source-backed routing registry, with the current Codex LLM by default or optional direct Jev assessment. Use when asked to route coding tasks, choose subagent models, or set up MiMo in Codex.٢٢ سبتمبر ٢٠٢٦
#٢٠codejunkie99/image-reconstruction١ skillsimage-reconstructionAnalyze image references into named visual elements and reconstruction JSON, then compile prompts for recreating, remixing, or selectively editing them. Use when the user wants to understand a reference, reuse its visual decisions, or preserve specific parts…١٢ سبتمبر ٢٠٢٦
#٢١codejunkie99/avids-essential-skills١ skillsclaude-code-hygieneAudit and enforce 7 Claude Code best practices that most users get wrong. Checks for CLAUDE.md quality, session hygiene, feedback loops, worktree usage, permission config, prompt specificity, and parallel session opportunities. Use when: starting a new…٦ أبريل ٢٠٢٦
#٢٢codejunkie99/dots-agent-team١ skillsdots-agent-teamCoordinate a dot-led team of separate CLI model workers through Model Router, using Jev for bounded routing, local memory actions and computer action choices. Use when asked to set up a multi-model agent team, resume approved project work, detect unfinished…٣٠ سبتمبر ٢٠٢٦