Coder/deepseek-harness
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(mode): an approved exit leaves plan mode at the step boundary, not mid-batch

Browse Source 
Review finding: the exit tool's direct mode/set append flipped the
folded mode while the loop could still execute further tool calls from
the SAME assistant response — a same-batch exit_plan_mode + write pair
would sail past tools/pre-execute under 'default' even though the
request was assembled under the plan-shaped header. That broke the
design's own invariant (a step's executions run under the mode its
assembly folded), which the pending-intent flush was built to hold for
user flips.

The tool now records the switch as a pending intent like every other
writer, flushed at this step's end (still in-turn); pending intents
carry a narrate flag so the exit's flush stays silent — the tool result
is its narration — while user flips keep the coalesced boundary notice.
The gate, folding the logged mode only, now provably covers the whole
batch: regression test pins approve-then-write-in-the-same-batch as
denied, and the widened toolset still arrives on the next step.

The recorded scenarios are re-recorded: the fixture now shows mode/set
landing after step/end, before the widened fallback header.
This commit is contained in:
kingwl
2026-07-10 10:35:43 +08:00
parent 4e09fdeafb
commit 14f7976e3d
11 changed files with 1122 additions and 838 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/rfc/implemented/feature/2026-07-07-plan-mode.md
View File

@@ -111,7 +111,7 @@ The gate folds the LOGGED mode only, never the pending intent — enforcement ju
### `exit_plan_mode`
`defineTool` with one required `plan: string` argument — the plan is thereby a durable, replayable log artifact riding the ordinary `tool/call` event. `execute` rejects an agent-less call (the [`todo_write` precedent](../../implemented/feature/2026-06-29-todo-write-tool.md)), re-checks the folded mode as defense in depth, then conducts the review: one single-select `ctx.userInteraction.ask()` question — approve, or keep planning — with the free-text channel open for feedback. Approve appends `mode/set { mode: 'default' }` in-turn (a tool execution is already turn-enclosed — the `todo/write` path) and returns a short confirmation; the next step's assembly restores the full toolset and logs the widening header event. Every other outcome — keep-planning (the user's feedback text carried verbatim), an aborted question, a missing provider — returns the corrective `isError` that tells the model to revise and re-present, and the mode stays `plan`.
`defineTool` with one required `plan: string` argument — the plan is thereby a durable, replayable log artifact riding the ordinary `tool/call` event. `execute` rejects an agent-less call (the [`todo_write` precedent](../../implemented/feature/2026-06-29-todo-write-tool.md)), re-checks the folded mode as defense in depth, then conducts the review: one single-select `ctx.userInteraction.ask()` question — approve, or keep planning — with the free-text channel open for feedback. Approve records the switch back to `default` as a SILENT boundary-applied pending intent (flushed at this step's end, still in-turn) and returns a short confirmation; the gate therefore stays plan-mode for every remaining call of the SAME assistant response — a same-batch `exit_plan_mode` + `write` pair cannot smuggle the write past a request assembled under the plan header — and the next step's assembly restores the full toolset and logs the widening header event. Every other outcome — keep-planning (the user's feedback text carried verbatim), an aborted question, a missing provider — returns the corrective `isError` that tells the model to revise and re-present, and the mode stays `plan`.
Its [render intent](../../implemented/architecture/2026-07-02-tool-render-intent-union.md), decided up front: `presentCall` is a `generic` card titled by the plan's first heading with the plan markdown as content, plus a `generic` result card — the review question arrives beside this already-streamed card, so what the human reviews is exactly the logged artifact. The seam is consumed opportunistically (`ctx.get('userInteraction')`), so `dsh-mode` composes without it and degrades to the manual exit pinned in the [FAQ](#faq).
@@ -131,7 +131,7 @@ No new cordis event is declared (`mode/set` rides `session/event`; the listeners
Each behind its own decision: subagent mode inheritance via a forwarded `AgentOptions.mode` (the option field itself ships), per-tool `ask` policies inside mode definitions (an OpenCode-style "bash asks in plan mode"), preset modes beyond `plan` (read-only, accept-edits), sandbox-backed bash confinement in plan mode, and the idle-record primitive if pending-intent loss proves real.
The recorded snapshot scenarios are landed: `plan-mode` (the pinned-header arc — plan-shaped initial header, scripted elicitation approve, the in-turn flip and widened fallback header, a real edit) and `plan-mode-reject` (keep-planning feedback carried verbatim in the corrective `isError`), beside the keyless `modes-advertise` wire golden. The gate's deny path stays pinned at the unit tier — the recorded model never calls a filtered tool, which is the behavior the soft layer exists to produce.
The recorded snapshot scenarios are landed: `plan-mode` (the pinned-header arc — plan-shaped initial header, scripted elicitation approve, the boundary-flushed flip and widened fallback header, a real edit) and `plan-mode-reject` (keep-planning feedback carried verbatim in the corrective `isError`), beside the keyless `modes-advertise` wire golden. The gate's deny path stays pinned at the unit tier — the recorded model never calls a filtered tool, which is the behavior the soft layer exists to produce.
## FAQ

4
docs/tool-catalog.md
View File

@@ -17,7 +17,7 @@ This table connects model-visible tool names to the plugin package and service s
| --- | --- | --- | --- | --- | --- |
| `@deepseek-ai/dsh-tool-ask-user` | `ask_user_question` | `ctx.tools`, `ctx.userInteraction` | `tool/call`, `tool/result after a UI/provider answers the question` | - | ask_user_question pauses the tool call until the active UI provider returns a human answer. |
| `@deepseek-ai/dsh-tools` | `run_code` | `ctx.tools`, `ctx.codeRuntime (execution time)`, `ctx.systemPrompt` | `tool/call`, `one tool/code-dispatch per bridged sub-call`, `tool/result` | - | Registered by the tool registry itself under `mode: code` / `mode: both` (see the Code Mode RFC). Under `code` it is the ONLY wire tool; the other registered tools are declared to the model as a generated TypeScript SDK prompt section instead, and a program calls them through port-bridged bindings that dispatch through the ordinary tools/pre-execute → tools/post-execute pipeline, one at a time. |
| `@deepseek-ai/dsh-mode` | `exit_plan_mode` | `ctx.tools`, `ctx.systemPrompt`, `ctx.userInteraction (execution time, opportunistic)` | `tool/call`, `mode/set back to default on an approved review`, `tool/result` | - | exit_plan_mode presents the plan for the user's review over the user-interaction seam (approve / keep planning with feedback); approval flips the logged session mode back to default. The assemble filter shows it only while the folded mode is plan. |
| `@deepseek-ai/dsh-mode` | `exit_plan_mode` | `ctx.tools`, `ctx.systemPrompt`, `ctx.userInteraction (execution time, opportunistic)` | `tool/call`, `mode/set back to default on an approved review`, `tool/result` | - | exit_plan_mode presents the plan for the user's review over the user-interaction seam (approve / keep planning with feedback); approval flips the logged session mode back to default at the step boundary (same-batch calls stay plan-gated). The assemble filter shows it only while the folded mode is plan. |
| `@deepseek-ai/dsh-tool-bash` | `bash`, `bash_kill`, `bash_output` | `ctx.tools`, `ctx.bash` | `tool/call`, `tool/result`, `context/message via agent.inject() for background completion notices` | - | The bash/bash_output/bash_kill tools are model-facing consumers of the bash executor seam. |
| `@deepseek-ai/dsh-tool-cordis` | `cordis_inspect`, `cordis_mount`, `cordis_unmount` | `ctx.tools` | `tool/call`, `tool/result`, `live plugin-tree mutations (mount/unmount)` | - | Ships in examples/cordis-agent only (a deliberate opt-in — mounted code gets the real ctx, see docs/rfc/implemented/feature/2026-07-08-self-referential-cordis-toolset.md). Plugins the model mounts may register ADDITIONAL model-visible tools at runtime; the request-header ToolsDelta logs those tool-set changes. |
| `@deepseek-ai/dsh-tool-fs` | `edit`, `read`, `write` | `ctx.tools`, `ctx.fs`, `ctx.systemPrompt` | `tool/call`, `fs/write-intent or fs/edit-intent for mutations`, `fs/observed after successful file operations`, `tool/result` | - | The read-before-write/edit policy is added by `@deepseek-ai/dsh-fs-policy` (an `fs/*` event-gate plugin, no schema change); a deployment that loads these tools is expected to also load it. The tool schemas above are identical with or without the policy plugin. |
@@ -143,7 +143,7 @@ Present your plan for the user's review and, on approval, leave plan mode. Send
Source: [`packages/mode/mode/src/index.ts`](../packages/mode/mode/src/index.ts)
exit_plan_mode presents the plan for the user's review over the user-interaction seam (approve / keep planning with feedback); approval flips the logged session mode back to default. The assemble filter shows it only while the folded mode is plan.
exit_plan_mode presents the plan for the user's review over the user-interaction seam (approve / keep planning with feedback); approval flips the logged session mode back to default at the step boundary (same-batch calls stay plan-gated). The assemble filter shows it only while the folded mode is plan.
## `@deepseek-ai/dsh-tool-bash`