Merge remote-tracking branch 'origin/master' into worktree-dynamic-workflows

Master's reconstructable-requests overhaul (#179) meets the workflow tool:
- subagent-inprocess structured-output nudge becomes a system-prompt section
  plus logged context (the injected-request waterfall shape is gone upstream)
- snapshot fixtures re-recorded on the merged tree so every request/header
  carries the workflow tool; authored error-finish/cancel headers patched to
  the merged tool list and system text
- architecture.md condensed back under its word ceiling; module graph regenerated
This commit is contained in:
Tianyi Cui
2026-07-06 22:28:57 +08:00
123 changed files with 7335 additions and 7188 deletions

View File

@@ -1,6 +1,6 @@
# DeepSeek Harness Architecture
The **DeepSeek Harness SDK** is an SDK for building agent harnesses using the Cordis framework. The governing principle is simple: **everything is a plugin**. For example, the shipped agent loop is just one plugin in the default bundle, not a privileged kernel.
The **DeepSeek Harness SDK** is a Cordis-based SDK for building agent harnesses. The governing principle: **everything is a plugin**; the shipped agent loop is just one plugin in the default bundle, not a privileged kernel.
Read this page as the system map before changing `packages/`. It explains how the runtime is shaped, how the default loop moves work, where state lives, and where extensions attach. Type shapes live in [core-data-structures/](core-data-structures/core.md); exact event and service signatures live in the generated [events](cordis-catalog/events.md) and [services](cordis-catalog/services.md) catalogs; package contracts live in the [package map](../packages/README.md); rationale lives in the [RFCs](rfc/README.md). If Cordis itself is new to you, start with the [Cordis primer](cordis-primer.md).
@@ -8,7 +8,7 @@ Read this page as the system map before changing `packages/`. It explains how th
A running harness is one Cordis context. Packages contribute service keys, typed events, and disposable registrations to that context. Services are the stable call surfaces (`ctx.llm`, `ctx.tools`, `ctx.sessions`); events are interception and notification points (`agent/request`, `tools/pre-execute`, `session/event`); registrations install prompt sections, tool schemas, providers, adapters, and listeners.
The default distribution is a composition, not a hierarchy. `packages/core/` is a repository grouping for the default agent spine; capability seams around it are equally first-class plugins from a Cordis perspective.
The default distribution is a composition, not a hierarchy. `packages/core/` is a repository grouping for the default agent spine; capability seams around it are equally first-class plugins to Cordis.
### Default Service Spine
@@ -42,7 +42,7 @@ Events are the harness extension API. Each service owns the vocabulary for the b
Use the event domain to decide where new behavior belongs:
- **Session events** are durable, replayable facts. Turn and step boundaries, user input, assistant output, tool calls, tool results, steering, compaction records, and tool-owned durable facts append to the session log and flow through `session/event`.
- **Agent events** are live runtime surfaces. They carry the live `Agent` handle for status, diagnostics, prompt admission, request mutation, result validation, and continuation policy.
- **Agent events** are live runtime surfaces. They carry the live `Agent` handle for status, diagnostics, prompt admission, call-config shaping, result validation, and continuation policy.
- **Capability events** belong to the seam that owns the action. `tools/*`, `llm/*`, `system-prompt/*`, `fs/*`, and `subagent/*` let policy and adapters attach without importing the loop.
### Interception Semantics
@@ -72,8 +72,8 @@ forever:
assemble system prompt and tool schemas
agent/pre-step
'step/start'
derive messages from the session log
agent/request -> llm/stream
snapshot the derived messages (the reconstruction boundary)
agent/request (config only) -> log request/header -> llm/stream (frozen)
'assistant/chunk'
agent/step-result
'assistant/message'
@@ -89,7 +89,7 @@ forever:
checkpoint persistence and notify idle/running status
```
Prompt assembly is single-path: `renderPrompt(assemble({ agent }))` IS the system prompt sent to the model. Plugins contribute ordered sections (static or computed from the per-call `AssembleContext`), tool schemas, and named variables interpolated as `{{name}}` at render — strictly, so an unknown or valueless reference fails the turn instead of shipping a hole. `dsh-system-prompt` itself owns the openers — the static `harness:identity` section (order 100) and the deployment's persona (order 0, from its `persona` config, shared by every agent in the context) — while the shipped loop registers the `model`/`cwd` variables; prompt-fact ownership is pinned by the [prompt-variables RFC](rfc/implemented/architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.md).
Prompt assembly is single-path: `renderPrompt(assemble({ agent }))` IS the system prompt sent to the model. Plugins contribute ordered sections (static or computed from the per-call `AssembleContext`), tool schemas, and named variables interpolated as `{{name}}` at render — strictly, so an unknown or valueless reference fails the turn instead of shipping a hole. `dsh-system-prompt` owns the openers — the static `harness:identity` section (order 100) and the deployment's persona (order 0, from its `persona` config, shared by every agent in the context) — while the shipped loop registers the `model`/`cwd` variables; prompt-fact ownership is pinned by the [prompt-variables RFC](rfc/implemented/architecture/2026-07-05-prompt-variables-and-tool-guidance-ownership.md).
Post-tool context lands after all tool results so tool-call/result adjacency stays stable. Steering drains between steps; leftover steering after a turn is re-queued as ordinary input.
@@ -109,6 +109,8 @@ Every session event is turn-enclosed. Reloading a crashed session preserves the
The session log is the source of truth. `deriveMessages()` projects session events into the `Message[]` sent to the model; raw `assistant/chunk` events stay in the log for replay and UI fidelity. Replay, fork, resume, transcript rendering, telemetry, and persistence all derive from the same event stream.
**Model-visible ⟺ logged**: the log reconstructs every conversation request byte-for-byte — messages by derivation at the `step/start` boundary, the header (system prompt, tools, model + sampling) by folding `request/header` events — asserted per request by the dev invariant ([reconstructability RFC](rfc/implemented/architecture/2026-07-05-reconstructable-requests.md)).
Durability is a plugin concern. Persistence backends buffer synchronous `session/event` notifications and the loop awaits a turn-end checkpoint before moving on. The `SessionPersistence` seam stores `SessionEvent` directly, with metadata in `SessionHeader`; JSONL and SQLite share one contract suite.
### Model Content