ZiyaZhang
e8f5c0b51b
Minimal Electron shell over the DSH JSON-RPC runtime — a first-look at what a ChatGPT.app-style host on top of the DeepSeek Harness looks like, with the harness's normally-invisible internals (trace timeline, context surface, subagent tree, compaction, plugin registry, rubrics) brought forward as first-class UI surfaces so plugin authors and researchers can see what the agent is actually doing. Runs against three keyless-to-live profiles (stdio-echo works on master out of the box; daemon-echo / daemon-vibe-echo activate once the daemon-demo lands; stdio-deepseek and daemon-vibe hit the real DeepSeek API when you supply a key). HARNESS_DEV auto-resolves to the in-repo runtime when this shell ships under examples/desktop/, so a fresh clone launches without config; env DSH_DEV_ROOT overrides for custom layouts, and a sibling deepseek-harness-dev/ checkout is the original dev workflow. Cold-clone gate (P0 fixes for first-time-clone usability): - HARNESS_DEV: 3-candidate resolver (env → walk-up in-repo marker → sibling), unit-tested via mock fs so ordering is locked without needing either real layout on disk. - config yml leaves rewritten at assemble time so the sibling-clone paths (../../deepseek-harness-dev/examples/echo-agent/…) become the in-repo paths (../../echo-agent/…) in the released tree — source yml stays usable for local dev, released tree ships a working shape. - pnpm-workspace.yaml allowBuilds.electron = true (was placeholder). - missing-key card in stdio-deepseek offers a one-click switch to stdio-echo (the keyless profile that works on master) rather than daemon-echo (blocked on the not-yet-shipped daemon-demo). - assemble-oss-release.sh rewrites the source-side breadcrumb name 'dsh-desktop-demo' → 'dsh-desktop' for the released package.json. FOUC guard on the onboarding gate (41fc5df carried) keeps the first-launch splash from flashing before the runtime probe finishes. Test suite (1634 tests in source, 3990 in the runtime repo) covers resolver ordering, renderer classifiers, trace timeline shape, compaction diff rendering, rubric parity, and the missing-key onboarding paths.
77 lines
4.3 KiB
JSON
77 lines
4.3 KiB
JSON
[
|
|
{
|
|
"_mock": true,
|
|
"_mockReason": "#162 rec 21 companion fixture (team-lead 2026-07-17 real-audience directive): a real research scenario where an engineer is debugging 'did my API/harness return thinking tokens or not?'. Two consecutive turns, SAME prompt shape, differing only in whether the assistant/message content carries a `reasoning` block (i.e., whether the request/header's request round-tripped a thinking field). Turn A has reasoning present (streamed via reasoning-delta chunks); turn B has no reasoning at all — same tokens output otherwise. Opening the reasoning block on A and finding no block on B gives a one-minute triage: 'the model DID think but the harness dropped it' vs 'the model returned no thinking' becomes visually obvious without diffing raw response bodies. The reasoning-tokens usage field is present on A, absent on B, mirroring the wire signal a responder would look at."
|
|
},
|
|
|
|
{ "type": "user/message", "time": 1721401800050, "seq": 801,
|
|
"data": { "content": [ { "type": "text", "text": "Turn A: what does dsh-brand's `pickPreset` fall back to when the preset name is unknown?" } ] } },
|
|
{ "type": "step/start", "time": 1721401800100, "seq": 802,
|
|
"data": { "turn": 8, "step": 0 } },
|
|
{ "type": "request/header", "time": 1721401800150, "seq": 803,
|
|
"data": {
|
|
"header": {
|
|
"model": "deepseek-chat",
|
|
"system": "You are a research coding agent.",
|
|
"tools": [ { "name": "read", "description": "Read a file." } ],
|
|
"config": { "temperature": 0.2, "maxTokens": 4096, "reasoning": { "enabled": true } }
|
|
},
|
|
"reason": "step-start"
|
|
}
|
|
},
|
|
|
|
{ "type": "assistant/chunk", "time": 1721401800200, "seq": 804,
|
|
"data": { "chunk": { "type": "reasoning-delta", "text": "The user is asking about pickPreset's unknown-name fallback." } } },
|
|
{ "type": "assistant/chunk", "time": 1721401800210, "seq": 805,
|
|
"data": { "chunk": { "type": "reasoning-delta", "text": " Looking at the code I remember it defaults to 'default' when the map lookup misses." } } },
|
|
{ "type": "assistant/chunk", "time": 1721401800220, "seq": 806,
|
|
"data": { "chunk": { "type": "reasoning-delta", "text": " Answering directly since we don't need a tool call to confirm this." } } },
|
|
|
|
{ "type": "assistant/chunk", "time": 1721401800240, "seq": 807,
|
|
"data": { "chunk": { "type": "text-delta", "text": "Falls back to the 'default' preset." } } },
|
|
|
|
{ "type": "assistant/message", "time": 1721401800300, "seq": 808,
|
|
"data": {
|
|
"content": [
|
|
{ "type": "reasoning", "text": "The user is asking about pickPreset's unknown-name fallback. Looking at the code I remember it defaults to 'default' when the map lookup misses. Answering directly since we don't need a tool call to confirm this." },
|
|
{ "type": "text", "text": "Falls back to the 'default' preset." }
|
|
],
|
|
"usage": { "inputTokens": 380, "outputTokens": 12, "reasoningTokens": 46 }
|
|
}
|
|
},
|
|
|
|
{ "type": "step/end", "time": 1721401800400, "seq": 809,
|
|
"data": { "turn": 8, "step": 0, "reason": "stop" } },
|
|
|
|
{ "type": "user/message", "time": 1721401801050, "seq": 810,
|
|
"data": { "content": [ { "type": "text", "text": "Turn B (rerun with reasoning disabled): what does dsh-brand's `pickPreset` fall back to when the preset name is unknown?" } ] } },
|
|
{ "type": "step/start", "time": 1721401801100, "seq": 811,
|
|
"data": { "turn": 9, "step": 0 } },
|
|
{ "type": "request/header", "time": 1721401801150, "seq": 812,
|
|
"data": {
|
|
"header": {
|
|
"model": "deepseek-chat",
|
|
"system": "You are a research coding agent.",
|
|
"tools": [ { "name": "read", "description": "Read a file." } ],
|
|
"config": { "temperature": 0.2, "maxTokens": 4096, "reasoning": { "enabled": false } }
|
|
},
|
|
"reason": "step-start"
|
|
}
|
|
},
|
|
|
|
{ "type": "assistant/chunk", "time": 1721401801240, "seq": 813,
|
|
"data": { "chunk": { "type": "text-delta", "text": "Falls back to the 'default' preset." } } },
|
|
|
|
{ "type": "assistant/message", "time": 1721401801300, "seq": 814,
|
|
"data": {
|
|
"content": [
|
|
{ "type": "text", "text": "Falls back to the 'default' preset." }
|
|
],
|
|
"usage": { "inputTokens": 380, "outputTokens": 12 }
|
|
}
|
|
},
|
|
|
|
{ "type": "step/end", "time": 1721401801400, "seq": 815,
|
|
"data": { "turn": 9, "step": 0, "reason": "stop" } }
|
|
]
|