Install any skill in seconds. Free to start, no credit card required.
Get Started Free →contextosai/contextos-architect
Design or assess an agent system using ContextOS's five-plane runtime model, canonical contracts, and staged adoption path.
Passed 16 of 21 test cases with the skill, and 9 of 21 without it.
and 8 more in the full report
8 fixed9 passed either way5 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 7 test cases the skill fixed. Same model, same prompt, with and without the skill.
We are designing an AI clinical triage assistant that evaluates patient symptoms and routes urgent cases to hospital ER staff via EHR APIs. We plan to tag this initial architecture spec as 'Phase 1 Draft Stage'. Please generate the complete architectural specification for this pipeline, including autonomy boundaries and plane allocations.
| Checked | Without skill | With skill |
|---|---|---|
| Explicitly declares target maturity as regulated/high-risk, prototype, controlled beta, or production. | Missed | Passed |
| Structures system responsibilities across Intelligence, Context, Decision, Action, and Trust planes. | Missed | Passed |
The instructions your model reads when the skill loads.
Turn a workflow into a governed decision runtime whose claims and effects are reconstructable. Preserve the user's product and stack choices; ContextOS supplies control boundaries, not a mandatory framework.
AGENTS.md and the current overview, architecture, governance, harness-engineering, API-contract, and Context Pack material. Prefer current schemas and code over remembered field shapes.Read references/architecture-map.md for the plane-by-plane questions, minimum artifacts, and rollout gates.
Define the governed outcome before choosing prompts or agents:
propose, recommend, execute, or escalate);Map the slice across all five planes:
Use typed seams—RunContext, ContextPack, CompiledContext, ToolEnvelope, DecisionSpec, DecisionRecord, and ReplayPacket—only where their current contracts are actually implemented. Do not invent canonical fields to make a diagram look complete.
test cases passed, vs 9 / 21 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.