| name | local-analysis-offload |
| description | Plan privacy-first local AI/model offload for expensive elder-care casework analysis. Use for Qwen/GLM/Gemma-class local models, LM Studio/lms, LM Studio OpenAI-compatible API, Goose CLI, Ollama, llama.cpp, vLLM, local VLM/image analysis, OCR, ffmpeg, Whisper, Parakeet, diarization, transcript enrichment, and batch prompt planning. Defaults to plan-only until the user approves running jobs. |
| argument-hint | Workload, input paths, output root, candidate tools/models, plan-only or run, validation standard |
Local Analysis Offload
This VS Code / Copilot skill wrapper points to the canonical Cowork / Claude Desktop plugin workflow.
Load these files
../../../elder-care-casework/skills/local-analysis-offload/SKILL.md../../../elder-care-casework/skills/audio-transcript-review/LOCAL_AUDIO_OFFLOAD.md for audio/transcript work.- Any workload-specific canonical skill, such as
../../../elder-care-casework/skills/snf-records-review/SKILL.md or ../../../elder-care-casework/skills/audio-transcript-review/SKILL.md.
Use this wrapper when
- The user wants to use local models or deterministic local tools before consuming main-agent context.
- The task is expensive, repetitive, or privacy-sensitive: OCR/image labeling, transcript cleanup, speaker attribution, table extraction, first-pass tagging, or packet review triage.
- The user mentions Qwen, GLM, Gemma, LM Studio,
lms, LM Studio OpenAI-compatible API, Goose, Ollama, llama.cpp, vLLM, ffmpeg, Whisper, Parakeet, diarization, VLM, local image analysis, or GPU offload.
Default to plan-only. Do not run local tools or models until the user explicitly approves the run.
