grid-operations
FeaturedLook after the models on the user's machines — one laptop or a fleet: recognise their private grid, pick the model and settings that fit this machine for coding without asking, start it with the longest context the machine can give (never under 64K), say what it costs in memory, prove it answers with one bounded call, change or stop a running one, and use Grid routing, usage, media and training commands.
Install
Quality Score: 93/100
Skill Content
Details
- Author
- autonomous-ai
- Repository
- autonomous-ai/openharness
- Created
- 1 months ago
- Last Updated
- today
- Language
- C
- License
- MIT
Similar Skills
Semantically similar based on skill content — not just same category
eval-bench
Run and read the local-model benchmark (`tools/local_model_eval.py`) against this server - the grid's fences, what the grader's matching rules are for, `--verify-key`, `--compare`/`--series`, and how a moved aggregate gets misread as a controlled result. Use when running the eval, adding or re-grading a task, or writing up a benchmark result.
gridhand
Interact with the desktop GUI — take screenshots, list/raise windows, click with grid targeting, type text, press key combos. Use when you need to see the screen, find windows, click on things, type into applications, or automate any GUI interaction. All commands return JSON.
agent-fleet
Act as an orchestrator that plans work, delegates each piece to the right model across Claude Code and Codex (Opus, Sol, Sonnet, Terra, Luna, Haiku, Mini, Spark), reviews what comes back, and escalates on failure. Use when the user wants work split across models or backends, wants a cheaper/faster model to do the grunt work, wants a second model to review, or says things like "delegate this", "fan this out", "use the fleet", "have Codex do X", "have Claude do X", "route this to the right model", or "orchestrate this".