Answers from the documentation; checking which model is on duty
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.
autoresearch - Skill - OpenSmartRoute
Skillv1.0.0
autoresearch
Stateful single-mission improvement loop with strict evaluator contract, markdown decision logs, and max-runtime stop behavior
Imported from yeachan-heo/oh-my-claudecode (skills/autoresearch/SKILL.md) via skills.sh. Install upstream with npx skills add yeachan-heo/oh-my-claudecode --skill autoresearch. Copyright stays with the author.
<Use_When>
You already have a mission and evaluator from /deep-interview --autoresearch
You want persistent single-mission improvement with strict evaluation
You need durable experiment logs under .omc/autoresearch/
You want a supported path for periodic reruns via Claude Code native cron
</Use_When>
<Do_Not_Use_When>
You need evaluator generation at runtime — use /deep-interview --autoresearch first
You need multiple missions orchestrated together — v1 forbids that
You want the deprecated omc autoresearch CLI flow — it is no longer authoritative
</Do_Not_Use_When>
<Required_Artifacts>
Canonical persistent storage lives under .omc/autoresearch/<mission-slug>/ and/or .omc/logs/autoresearch/<run-id>/.
Reuse existing runtime artifacts when available rather than duplicating them unnecessarily.
</Required_Artifacts>
<Cron_Integration>
Claude Code native cron is a supported integration point for periodic mission enhancement. In v1, prefer documenting/configuring cron inputs over building a large scheduler UI.
If cron is used:
keep one mission per scheduled job
preserve the same mission/evaluator contract
append new run artifacts rather than overwriting prior experiments
</Cron_Integration>
<Execution_Policy>
Do not hand execution back to omc autoresearch
Do not create multi-mission orchestration
Prefer reusing src/autoresearch/* runtime/schema helpers where they already match the stricter contract
Keep logs useful to humans, not only machines
</Execution_Policy>
Installed into a catalogue, chosen by a router
Install autoresearch 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.
When a provider is down or over quota the request moves on; the trace says it did.
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 slotcurl-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}'
Manifest
An Open Capability Manifest: the router reads it to know what this does, what it costs and when to pick it.
{
"ocm": "1",
"id": "yeachan-heo-oh-my-claudecode-autoresearch",
"kind": "skill",
"name": "autoresearch",
"description": "Stateful single-mission improvement loop with strict evaluator contract, markdown decision logs, and max-runtime stop behavior",
"publisher": "yeachan-heo",
"version": "1.0.0",
"capabilities": {
"domains": [
"observability",
"legal"
],
"tags": [
"skill-md",
"skills-sh"
],
"languages": [
"en"
]
},
"quality_prior": 0.6,
"examples": [
"Stateful single-mission improvement loop with strict evaluator contract, markdown decision logs, and max-runtime stop behavior"
],
"primary": false,
"metadata": {
"source": {
"provider": "skills.sh",
"repository": "https://github.com/yeachan-heo/oh-my-claudecode",
"path": "skills/autoresearch/SKILL.md",
"ref": "HEAD",
"url": "https://www.skills.sh/yeachan-heo/oh-my-claudecode/autoresearch",
"key": "yeachan-heo/oh-my-claudecode/skills/autoresearch/SKILL.md"
}
},
"instructions": "<Purpose>\nAutoresearch is a stateful skill for bounded, evaluator-driven iterative improvement. It owns one mission at a time, keeps iterating through non-passing results, records each evaluation and decision as durable artifacts, and stops only when an explicit max-runtime ceiling or another explicit terminal condition is reached.\n</Purpose>\n\n<Use_When>\n- You already have a mission and evaluator from `/deep-interview --autoresearch`\n- You want persistent single-mission improvement with strict evaluation\n- You need durable experiment logs under `.omc/autoresearch/`\n- You want a supported path for periodic reruns via Claude Code native cron\n</Use_When>\n\n<Do_Not_Use_When>\n- You need evaluator generation at runtime — use `/deep-interview --autoresearch` first\n- You need multiple missions orchestrated together — v1 forbids that\n- You want the deprecated `omc autoresearch` CLI flow — it is no longer authoritative\n</Do_Not_Use_When>\n\n<Contract>\n- Single-mission only in v1\n- Mission setup/evaluator generation stays in `deep-interview --autoresearch`\n- Evaluator output must be structured JSON with required boolean `pass` and optional numeric `score`\n- Non-passing iterations do **not** stop the run\n- Stop conditions are explicit and bounded, with max-runtime as the primary strict stop hook\n</Contract>\n\n<Required_Artifacts>\nCanonical persistent storage lives under `.omc/autoresearch/<mission-slug>/` and/or `.omc/logs/autoresearch/<run-id>/`.\n\nMinimum required artifacts:\n- mission spec\n- evaluator script or command reference\n- per-iteration evaluation JSON\n- markdown decision logs\n\nRecommended canonical shape:\n```text\n.omc/autoresearch/<mission-slug>/\n mission.md\n evaluator.json\n runs/<run-id>/\n evaluations/\n iteration-0001.json\n iteration-0002.json\n decision-log.md\n```\nReuse existing runtime artifacts when available rather than duplicating them unnecessarily.\n</Required_Artifacts>\n\n<Workflow>\n1. Confirm a single mission exists and evaluator setup is already available.\n2. Ensure mode/state is active for `autoresearch` and records:\n - mission slug/dir\n - evaluator reference\n - iteration count\n - started/updated timestamps\n - explicit max-runtime or deadline\n3. On every iteration:\n - run exactly one experiment/change cycle\n - run the evaluator\n - persist machine-readable evaluation JSON\n - append a human-readable markdown decision log entry\n - continue even when evaluation does not pass\n4. Stop when:\n - max-runtime ceiling is reached\n - user explicitly cancels\n - another explicit terminal condition is recorded by the runtime\n</Workflow>\n\n<Cron_Integration>\nClaude Code native cron is a supported integration point for periodic mission enhancement. In v1, prefer documenting/configuring cron inputs over building a large scheduler UI.\n\nIf cron is used:\n- keep one mission per scheduled job\n- preserve the same mission/evaluator contract\n- append new run artifacts rather than overwriting prior experiments\n</Cron_Integration>\n\n<Execution_Policy>\n- Do not hand execution back to `omc autoresearch`\n- Do not create multi-mission orchestration\n- Prefer reusing `src/autoresearch/*` runtime/schema helpers where they already match the stricter contract\n- Keep logs useful to humans, not only machines\n</Execution_Policy>",
"cost": {
"context_tokens": 823
}
}
Fetch it by URL: GET /api/v1/registry/yeachan-heo-oh-my-claudecode-autoresearch/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.
