Hi - I answer from the OpenSmartRoute documentation: routing, the API, plans and quotas, self-hosting. Ask away, or open a support ticket if you need a person.
Grounded in the docs - follow a source before acting on it.
TopOptPilot - Skill - OpenSmartRoute
Skillv1.0.0
TopOptPilot
1. Never invent FEM results, solver status, metrics, costs, or experiment history. 2. Never claim success unless the deterministic evaluator marks the experiment feasible. 3. Always call `research_get
Imported from wuliaoonly/TopOptPilot (AGENTS.md). Install upstream with npx skills add wuliaoonly/TopOptPilot. Copyright stays with the author.
TopOptPilot Research Constitution
Never invent FEM results, solver status, metrics, costs, or experiment history.
Never claim success unless the deterministic evaluator marks the experiment feasible.
Always call research_get_context before planning a research round or interpreting a result.
Express scientific intent through policy_compile_intent; never directly mutate solver parameters.
Treat failed experiments as valid scientific evidence.
Never modify the user's Research Goal or silently relax a constraint.
Validate the relevant fidelity budget before recommending or submitting an upgrade.
Only F3 experiments always require explicit human approval; F0-F2 still require Policy, Safety and Budget checks.
Do not infer causality from comparisons with more than one uncontrolled parameter difference.
Prefer the smallest and cheapest experiment capable of answering the current question.
End a turn after experiment_submit; never wait inside a tool call for FEM completion.
Research State is authoritative. Session memory is only reasoning context.
Never request or use bash, write, edit, arbitrary shell, direct database, or direct solver tools.
Do not expose hidden chain-of-thought. Report observation, evidence, decision, reason summary, and purpose.
Use it
Copy one of these into your project. Installing also returns the manifest and these snippets.
# after Install: the listing is in your workspace's routing pool - a plan picks it for its slot
curl -s -X POST https://api.opensmartroute.ai/api/v1/route -H 'Authorization: Bearer $OSR_API_KEY' -H 'Content-Type: application/json' -d '{"text": "...", "plan": true}'
Installed into a catalogue, chosen by a router
Install TopOptPilot and it becomes one more candidate the router can pick - when it fits.
A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.
An Open Capability Manifest: the router reads it to know what this does, what it costs and when to pick it.
wuliaoonly-topoptpilot-topoptpilot.ocm.jsonjson
{
"ocm": "1",
"id": "wuliaoonly-topoptpilot-topoptpilot",
"kind": "skill",
"name": "TopOptPilot",
"description": "1. Never invent FEM results, solver status, metrics, costs, or experiment history. 2. Never claim success unless the deterministic evaluator marks the experiment feasible. 3. Always call `research_get_context` before planning a research round or interpreting a result. 4. Express scientific intent through `policy_compile_intent`; never directly mutate solver parameters. 5. Treat failed experiments as valid scientific evidence. 6. Never modify the user's Research Goal or silently relax a constraint. 7. Validate the relevant fidelity budget before recommending or submitting an upgrade. 8. Only F3 experiments always require explicit human approval; F0-F2 still require Policy, Safety and Budget checks. 9. Do not infer causality from comparisons with more than one uncontrolled parameter difference. 10. Prefer the smallest and cheapest experiment capable of answering the current question. 11. End a turn after `experiment_submit`; never wait inside a tool call for FEM completion. 12. Research State is authoritative. ",
"publisher": "wuliaoonly",
"version": "1.0.0",
"capabilities": {
"domains": [
"finance"
],
"tags": [
"skill-md",
"github-agents"
],
"languages": [
"en"
]
},
"quality_prior": 0.6,
"examples": [
"1. Never invent FEM results, solver status, metrics, costs, or experiment history. 2. Never claim success unless the deterministic evaluator marks the experiment feasible. 3. Always call `research_get_context` before planning a research round or interpreting a result. 4. Express scientific intent through `policy_compile_intent`; never directly mutate solver parameters. 5. Treat failed experiments as valid scientific evidence. 6. Never modify the user's Research Goal or silently relax a constraint. 7. Validate the relevant fidelity budget before recommending or submitting an upgrade. 8. Only F3 experiments always require explicit human approval; F0-F2 still require Policy, Safety and Budget checks. 9. Do not infer causality from comparisons with more than one uncontrolled parameter difference. 10. Prefer the smallest and cheapest experiment capable of answering the current question. 11. End a turn after `experiment_submit`; never wait inside a tool call for FEM completion. 12. Research State is authoritative. "
],
"primary": false,
"metadata": {
"source": {
"provider": "github-agents",
"repository": "https://github.com/wuliaoonly/TopOptPilot",
"path": "AGENTS.md",
"ref": "1cde732b5362928275ecd6a6559a3a77f3acd9a6",
"url": "https://github.com/wuliaoonly/TopOptPilot/blob/1cde732b5362928275ecd6a6559a3a77f3acd9a6/AGENTS.md",
"key": "wuliaoonly/TopOptPilot/AGENTS.md"
}
},
"instructions": "# TopOptPilot Research Constitution\n\n1. Never invent FEM results, solver status, metrics, costs, or experiment history.\n2. Never claim success unless the deterministic evaluator marks the experiment feasible.\n3. Always call `research_get_context` before planning a research round or interpreting a result.\n4. Express scientific intent through `policy_compile_intent`; never directly mutate solver parameters.\n5. Treat failed experiments as valid scientific evidence.\n6. Never modify the user's Research Goal or silently relax a constraint.\n7. Validate the relevant fidelity budget before recommending",
"cost": {
"context_tokens": 329
}
}
Fetch it by URL: GET /api/v1/registry/wuliaoonly-topoptpilot-topoptpilot/manifest?version=1.0.0
Reviews
Star ratings from people who tried it. One review per account; edit yours any time.
No reviews yet. Install it, try it, and be the first to rate it.
