subprocess: one explicit env channel on the spawn spec

Drop SubprocessSpawnSpec.dshEnv and splitEnvChannels(); childEnv() is now
scrubbed-base + explicit entries with no namespace validation. The invariant
dropped is the reserved-namespace check on explicit entries (DSH_* rejected
from env, non-DSH_* rejected from dshEnv). Explicit-entry trust already
covers it: an explicit credential-shaped entry has always merged after the
scrub as a deliberate caller opt-in, and an explicit DSH_* entry is the same
deliberate act — the staleness invariant lives entirely in scrubbedParentEnv
dropping AMBIENT credential-shaped and DSH_* names, which stays. The
validation's only observed effect was rejecting legitimate explicit entries:
both recent CI breakages (DSH_GATE_CONCURRENCY exported into every job
crashing lsp specs, DSH_PERMISSION_MODE in acp config.env crashing the
child spawn) were this check firing on values a caller meant to pass, each
fixed by routing around the bureaucracy the seam itself imposed.

The bash seam keeps its own request/spec dshEnv field: that is bash-owned
trusted-plugin vocabulary (the ctx.bashEnv collected overlay) whose merge-last
position guarantees a caller env entry cannot displace a managed fact;
bash-local now flattens ENV_OVERRIDES -> spec.env -> spec.dshEnv into the
seam's one env map. subagent-acp and lsp-local pass their single config env
map straight through. DshEnvironment/DshEnvironmentKey/DSH_ENV_PREFIX stay on
the subprocess seam as the namespace vocabulary (bash re-exports them;
scrubbedParentEnv filters on the prefix).

Tests: the two channel-rejection specs and the splitEnvChannels partition
spec are deleted; one spawn spec now proves an explicit DSH_* env entry
reaches the child while an ambient one is scrubbed; the acp/lsp forwarding
specs keep their MOCK_ECHO_ENV / LSP_FAKE_ECHO_ENV assertions with the split
comments rewritten to merge-after-scrub. Docs (en+zh, re-recorded) and the
owning Agent Notes updated; cordis api/services catalogs regenerated.
This commit is contained in:
Tianyi Cui
2026-07-27 04:14:51 +08:00
parent cb1864795e
commit 5717726835
43 changed files with 134 additions and 201 deletions

View File

@@ -19,8 +19,8 @@ import { clampTimeout, deadline, timeoutOf } from '@deepseek-ai/dsh-timeout'
* Model-friendly environment overrides: disable colors, pagers, and
* interactive terminal features that would garble tool output (the same set
* Codex hardcodes; Claude Code achieves it via TERM=dumb). Bash-tool policy —
* merged into the ordinary env channel, so a trusted caller's own entry still
* wins; the subprocess service applies its credential scrub independently.
* merged first into the spawn's explicit env, so a trusted caller's own entry
* still wins; the subprocess service applies its credential scrub independently.
*/
export const ENV_OVERRIDES = {
NO_COLOR: '1',
@@ -152,8 +152,10 @@ export class LocalBashExecutor extends BashExecutor {
},
graceMs: this.config.graceMs,
signal,
env: { ...ENV_OVERRIDES, ...spec.env },
dshEnv: spec.dshEnv,
// One explicit env map for the seam, layered so the trusted dshEnv
// snapshot beats both the caller's env and the terminal overrides; the
// subprocess service merges the whole map after its ambient scrub.
env: { ...ENV_OVERRIDES, ...spec.env, ...spec.dshEnv },
}
}

View File

@@ -2,5 +2,5 @@
# side as of the last confirmed-consistent state. Both languages carry equal authority;
# after editing either side, bring the other along and re-record with:
# pnpm run verify-translation-pairing --write
README.md: b4ee66a1fa2696254a1f2f411b7db5d3190f8370
README.zh.md: 151d4bd7ab257234584b9008c96e6356d7e39351
README.md: d7bf746969f52000fe298b65b995b7c631d8001c
README.zh.md: 14476b770397e4ef850c7c867e3058e25085c339

View File

@@ -33,7 +33,7 @@ Implementations subclass `BashExecutor` and implement the abstract methods. Disp
The per-session sandbox-mode override vocabulary (the `'sandbox/mode'` event, the `effectiveSandboxMode(events)` fold, and the `setSandboxMode(session, mode)` write path) is NOT here — it is policy state shared by every enforcing family, owned by [`@deepseek-ai/dsh-sandbox-policy`](../../sandbox/sandbox-policy/). `run()` returns `BashRunResult`; `start()` returns `BashProcess`, whose incremental read and kill methods are adapted by `dsh-tool-bash` into a generic task registration. A sandboxing executor stamps `BashSandboxInfo` on foreground results and settled process handles. See `src/types.ts` and [core-data-structures/bash.md](../../../docs/core-data-structures/bash.md).
`stdin` and ordinary `env` are set by in-process plugins (the hooks bridges, native plugins) to feed a hook command its JSON payload and `CLAUDE_PROJECT_DIR`/`CLAUDE_PLUGIN_ROOT` values. `dshEnv` is a separate trusted overlay restricted by type to managed keys; the exported `DSH_ENV_PREFIX` is the single source for that namespace, its `DshEnvironmentKey` template type, executor scrubbing, registry validation, derived built-in names, and model guidance. Model bash uses the current snapshot collected by `ctx.bashEnv`. Implementations remove inherited managed keys, reject those names in ordinary `env`, then merge `dshEnv`, so an omitted current fact cannot fall back to stale ambient state. The model-facing tool exposes none of these as parameters. All three remain optional on the resolved spec; absent means no input/overlay. See [the bash-stdin-env Agent Note](../../../.agents/notes/implemented/architecture/2026-06-30-bash-stdin-env-trusted-plugin-surface.md) and [the session environment Agent Note](../../../.agents/notes/implemented/feature/2026-07-10-agent-session-identity-and-log-location.md).
`stdin` and ordinary `env` are set by in-process plugins (the hooks bridges, native plugins) to feed a hook command its JSON payload and `CLAUDE_PROJECT_DIR`/`CLAUDE_PLUGIN_ROOT` values. `dshEnv` is a separate trusted overlay restricted by type to managed keys; the exported `DSH_ENV_PREFIX` is the single source for that namespace, its `DshEnvironmentKey` template type, executor scrubbing, registry validation, derived built-in names, and model guidance. Model bash uses the current snapshot collected by `ctx.bashEnv`. Implementations remove inherited managed keys, then merge `dshEnv` after ordinary `env`, so an omitted current fact cannot fall back to stale ambient state and an `env` entry cannot displace a managed value. The model-facing tool exposes none of these as parameters. All three remain optional on the resolved spec; absent means no input/overlay. See [the bash-stdin-env Agent Note](../../../.agents/notes/implemented/architecture/2026-06-30-bash-stdin-env-trusted-plugin-surface.md) and [the session environment Agent Note](../../../.agents/notes/implemented/feature/2026-07-10-agent-session-identity-and-log-location.md).
## Model Experience

View File

@@ -33,7 +33,7 @@
每会话沙箱模式覆盖词汇(`'sandbox/mode'` 事件、`effectiveSandboxMode(events)` fold 以及 `setSandboxMode(session, mode)` 写入路径)不位于此处。它是所有强制执行家族共享的策略状态,属于 [`@deepseek-ai/dsh-sandbox-policy`](../../sandbox/sandbox-policy/)。`run()` 返回 `BashRunResult`;`start()` 返回 `BashProcess`,其增量读取与终止方法由 `dsh-tool-bash` 适配为通用任务注册。沙箱执行器会在前台结果与已结算进程句柄上标记 `BashSandboxInfo`。详见 `src/types.ts` 与 [core-data-structures/bash.md](../../../docs/core-data-structures/bash.md)。
`stdin` 与普通 `env` 由同进程插件(hooks 桥接、原生插件)设置,用于向 hook 命令提供其 JSON payload 和 `CLAUDE_PROJECT_DIR`/`CLAUDE_PLUGIN_ROOT` 值。`dshEnv` 是受类型限制、仅允许受管 key 的独立受信任 overlay;导出的 `DSH_ENV_PREFIX` 是该 namespace、其 `DshEnvironmentKey` 模板类型、执行器清理、注册表验证、派生内置名称与模型指引的单一真源。模型 bash 使用 `ctx.bashEnv` 收集的当前快照。实现会移除继承的受管 key,拒绝普通 `env` 中的这些名称,再合并 `dshEnv`,因此省略的当前事实不会回退到陈旧环境状态。面向模型的工具不公开任何一个字段。这三者在已解析 spec 上仍然可选;缺失表示没有输入/overlay。详见 [bash-stdin-env Agent Note](../../../.agents/notes/implemented/architecture/2026-06-30-bash-stdin-env-trusted-plugin-surface.md) 与 [会话环境 Agent Note](../../../.agents/notes/implemented/feature/2026-07-10-agent-session-identity-and-log-location.md)。
`stdin` 与普通 `env` 由同进程插件(hooks 桥接、原生插件)设置,用于向 hook 命令提供其 JSON payload 和 `CLAUDE_PROJECT_DIR`/`CLAUDE_PLUGIN_ROOT` 值。`dshEnv` 是受类型限制、仅允许受管 key 的独立受信任 overlay;导出的 `DSH_ENV_PREFIX` 是该 namespace、其 `DshEnvironmentKey` 模板类型、执行器清理、注册表验证、派生内置名称与模型指引的单一真源。模型 bash 使用 `ctx.bashEnv` 收集的当前快照。实现会移除继承的受管 key,再在普通 `env` 之后合并 `dshEnv`,因此省略的当前事实不会回退到陈旧环境状态,`env` 条目也无法顶掉受管值。面向模型的工具不公开任何一个字段。这三者在已解析 spec 上仍然可选;缺失表示没有输入/overlay。详见 [bash-stdin-env Agent Note](../../../.agents/notes/implemented/architecture/2026-06-30-bash-stdin-env-trusted-plugin-surface.md) 与 [会话环境 Agent Note](../../../.agents/notes/implemented/feature/2026-07-10-agent-session-identity-and-log-location.md)。
## 模型体验

View File

@@ -60,17 +60,18 @@ export interface BashExecRequest {
stdin?: string | undefined
/**
* Ordinary environment entries for the command, merged after the credential
* scrub. `DSH_*` is reserved for {@link dshEnv} and implementations reject it
* here. Set by in-process plugins (the hooks bridges set
* `CLAUDE_PROJECT_DIR`, `CLAUDE_PLUGIN_ROOT`, …); the model-facing bash tool
* does not expose it as a parameter.
* scrub. Managed facts belong in {@link dshEnv}, which merges after this
* map, so an entry here can never displace one. Set by in-process plugins
* (the hooks bridges set `CLAUDE_PROJECT_DIR`, `CLAUDE_PLUGIN_ROOT`, …); the
* model-facing bash tool does not expose it as a parameter.
*/
env?: Record<string, string> | undefined
/**
* Harness-owned `DSH_*` variables for this execution. Executors discard
* ambient `DSH_*` entries before merging this snapshot, so an unavailable
* current fact cannot inherit a stale value from the harness process, and
* reject non-`DSH_*` names supplied through this managed channel.
* Harness-owned `DSH_*` variables for this execution (typed to managed
* keys). Executors discard ambient `DSH_*` entries before merging this
* snapshot last, so an unavailable current fact cannot inherit a stale
* value from the harness process and a caller {@link env} entry cannot
* displace a managed one.
*/
dshEnv?: DshEnvironment | undefined
/** Fully resolved per-call sandbox policy; sandboxing executors default it. */
@@ -97,12 +98,12 @@ export interface BashExecSpec {
stdin?: string | undefined
/**
* Ordinary environment entries carried through from
* {@link BashExecRequest.env}. `DSH_*` remains reserved for {@link dshEnv}.
* {@link BashExecRequest.env}; {@link dshEnv} still merges after them.
* OPTIONAL on the spec for the same reason as `stdin`: absent means no
* ordinary extra environment.
*/
env?: Record<string, string> | undefined
/** Managed `DSH_*` snapshot; implementations reject ordinary names. */
/** Managed `DSH_*` snapshot (typed to managed keys); merges after {@link env}. */
dshEnv?: DshEnvironment | undefined
/** Resolved sandbox policy; ignored by executors that do not confine. */
sandboxPolicy: SandboxExecutionPolicy | undefined

View File

@@ -2247,7 +2247,7 @@ export const TYPE_API: readonly TypeApiEntry[] = [
},
{
name: 'SubprocessSpawnSpec',
declaration: 'export interface SubprocessSpawnSpec {\n argv: readonly string[];\n cwd: string;\n stdio: SubprocessStdio;\n graceMs: number;\n signal?: AbortSignal | undefined;\n env?: Record<string, string> | undefined;\n dshEnv?: DshEnvironment | undefined;\n}',
declaration: 'export interface SubprocessSpawnSpec {\n argv: readonly string[];\n cwd: string;\n stdio: SubprocessStdio;\n graceMs: number;\n signal?: AbortSignal | undefined;\n env?: Record<string, string> | undefined;\n}',
},
{
name: 'SubprocessStdinMode',

View File

@@ -2,5 +2,5 @@
# side as of the last confirmed-consistent state. Both languages carry equal authority;
# after editing either side, bring the other along and re-record with:
# pnpm run verify-translation-pairing --write
README.md: 3c70977b6ea783d5aa1766709d2a507e08b8ceca
README.zh.md: d24100c5f850172845a24acadcbd2d0207b64859
README.md: 85254eea2bb74df277be6fd5de1529b5da4ea178
README.zh.md: 290601d455ba2012d0c1f5c97503ae87304881cb

View File

@@ -23,7 +23,7 @@ The `servers` record key is the stable provider id reserved on `ctx.lsp`; each v
|---|---|---|
| `command` | (required) | Executable to spawn — absolute, or resolved on the child PATH at load. Launch uses no shell. |
| `args` | `[]` | Arguments passed to the executable. |
| `env` | `{}` | Extra env merged on top of the credential-scrubbed ambient env (vars matching `KEY`/`SECRET`/`TOKEN` are not forwarded); `DSH_*` entries ride the subprocess seam's managed channel. |
| `env` | `{}` | Extra env merged on top of the credential-scrubbed ambient env (vars matching `KEY`/`SECRET`/`TOKEN` are not forwarded); an explicit `DSH_*` entry merges after the seam's scrub of ambient ones. |
| `extensionToLanguage` | (required) | Lowercase leading-dot extension → LSP language id (e.g. `{ '.ts': 'typescript' }`). |
| `initializationOptions` | `null` | Static `initialize` options forwarded to the server. |
| `configuration` | `null` | Static answer to every `workspace/configuration` item. |

View File

@@ -23,7 +23,7 @@ Namespace 插件(`name`/`inject`/`Config`/`apply`,无默认导出)
|---|---|---|
| `command` | (必填) | 要 spawn 的可执行文件:绝对路径,或在加载时从子进程 PATH 解析。不使用 shell 启动。 |
| `args` | `[]` | 传给可执行文件的参数。 |
| `env` | `{}` | 合并到已清理 credential 的环境之上的额外 env(匹配 `KEY`/`SECRET`/`TOKEN` 的变量不会转发);`DSH_*` 条目走 subprocess seam 的受管通道。 |
| `env` | `{}` | 合并到已清理 credential 的环境之上的额外 env(匹配 `KEY`/`SECRET`/`TOKEN` 的变量不会转发);显式 `DSH_*` 条目在 seam 清除环境中同名值之后合并。 |
| `extensionToLanguage` | (必填) | 小写、以点开头的扩展名 → LSP language id(例如 `{ '.ts': 'typescript' }`)。 |
| `initializationOptions` | `null` | 转发给服务器的静态 `initialize` 选项。 |
| `configuration` | `null` | 每个 `workspace/configuration` 配置项的静态答案。 |

View File

@@ -11,7 +11,6 @@
*/
import type { Writable } from 'node:stream'
import { splitEnvChannels } from '@deepseek-ai/dsh-subprocess'
import type { SubprocessHandle, SubprocessSpawnSpec } from '@deepseek-ai/dsh-subprocess'
import { encodeMessage, MessageDecoder } from './framing.ts'
@@ -99,9 +98,10 @@ export class LspConnection {
stderr: { maxBytes: spec.maxStderrBytes },
},
graceMs: spec.pipeDrainGraceMs,
// spec.env mixes the scrubbed base with explicit config entries; a
// configured DSH_* fact takes the managed channel the seam reserves.
...splitEnvChannels(spec.env),
// spec.env mixes the scrubbed base with explicit config entries; the
// seam merges the whole map after its own ambient scrub, so a
// configured DSH_* fact reaches the child.
env: spec.env,
})
/* v8 ignore start -- 'pipe' dispositions expose both streams by the seam contract; defensive. */
if (this.handle.stdin === undefined || this.handle.stdout === undefined) {

View File

@@ -51,10 +51,10 @@ describe('LspConnection', () => {
expect(conn.pid).toBeGreaterThan(0)
})
it('routes explicit DSH_* env entries onto the managed channel', async () => {
// A configured DSH_* fact must reach the child: the ordinary channel
// rejects the reserved namespace, so the connection's spawn must split it
// onto dshEnv. The fixture echoes the named variable back as hover text.
it('forwards explicit DSH_* env entries to the child', async () => {
// A configured DSH_* fact must reach the child: the seam scrubs only the
// ambient namespace, and the explicit entry merges after that scrub. The
// fixture echoes the named variable back as hover text.
const conn = connect({ LSP_FAKE_ECHO_ENV: 'DSH_LSP_TEST_FACT', DSH_LSP_TEST_FACT: 'managed' })
await conn.request('initialize', { capabilities: {} })
expect(await conn.request('textDocument/hover', {})).toEqual({ contents: 'managed' })

View File

@@ -2,5 +2,5 @@
# side as of the last confirmed-consistent state. Both languages carry equal authority;
# after editing either side, bring the other along and re-record with:
# pnpm run verify-translation-pairing --write
README.md: 0c21e179bd5ca405057b682786c41c8447a78317
README.zh.md: b349443cfcda80caa9b06cfa1cf80f6e780a98be
README.md: 044a4c5a71d32224262acef9e2c849f375198c2c
README.zh.md: 918df12ea0ea50e6f59b144e74601ff6beacbe86

View File

@@ -57,7 +57,7 @@ ACP advertises no start-time capabilities because this process cannot enforce th
## Process boundary
The child spawns through the [`dsh-subprocess`](../../subprocess/subprocess/README.md) seam: credential-shaped ambient variables are removed by the shared scrub, then explicit `config.env` values merge after it (an intended `DEEPSEEK_API_KEY` survives; `DSH_*` entries such as `DSH_PERMISSION_MODE` ride the seam's managed channel, since the scrubbed base reserves that namespace for current deployment facts), stderr is inherited to the parent's own stream, and disposal runs the seam's cooperative stdin-EOF→SIGTERM→SIGKILL ladder with this plugin's configured graces. The ACP wire is the real serialization boundary; same-process subagent values are not defensively cloned.
The child spawns through the [`dsh-subprocess`](../../subprocess/subprocess/README.md) seam: credential-shaped ambient variables and ambient `DSH_*` names are removed by the shared scrub, then explicit `config.env` values merge after it (an intended `DEEPSEEK_API_KEY` survives, and a `DSH_*` deployment fact such as `DSH_PERMISSION_MODE` reaches the child the same way — the scrub drops only its stale ambient namesake), stderr is inherited to the parent's own stream, and disposal runs the seam's cooperative stdin-EOF→SIGTERM→SIGKILL ladder with this plugin's configured graces. The ACP wire is the real serialization boundary; same-process subagent values are not defensively cloned.
The package has no default export. Cordis loader unwrapping would otherwise hide the named `inject` metadata; see [postmortem 0001](../../../docs/postmortem/0001-acp-default-export-drops-inject.md).

View File

@@ -57,7 +57,7 @@ ACP 不声明任何启动时能力,因为当前进程无法强制执行远程
## 进程边界
子进程经由 [`dsh-subprocess`](../../subprocess/subprocess/README.md) seam spawn:共享的凭据清除先移除名称形似凭据的环境变量,显式 `config.env` 值在清除之后合并(有意转发的 `DEEPSEEK_API_KEY` 会保留下来;`DSH_PERMISSION_MODE` 这类 `DSH_*` 条目走该 seam 的受管通道,因为清除后的基底把这一命名空间保留给当前部署事实),stderr 以 inherit 方式直通父进程自身的流,dispose 则以本插件配置的宽限期运行该 seam 的协作式 stdin EOF→SIGTERM→SIGKILL 阶梯。ACP 协议是真正的序列化边界;同进程 subagent 值不会为防御目的而克隆。
子进程经由 [`dsh-subprocess`](../../subprocess/subprocess/README.md) seam spawn:共享的凭据清除先移除名称形似凭据的环境变量和环境中已有的 `DSH_*` 名称,显式 `config.env` 值在清除之后合并(有意转发的 `DEEPSEEK_API_KEY` 会保留下来,`DSH_PERMISSION_MODE` 这类 `DSH_*` 部署事实也以同样的方式到达子进程——清除只丢弃其陈旧的同名环境值),stderr 以 inherit 方式直通父进程自身的流,dispose 则以本插件配置的宽限期运行该 seam 的协作式 stdin EOF→SIGTERM→SIGKILL 阶梯。ACP 协议是真正的序列化边界;同进程 subagent 值不会为防御目的而克隆。
本包没有默认导出。否则 Cordis loader 的解包会隐藏具名 `inject` 元数据;见[事故复盘 0001](../../../docs/postmortem/0001-acp-default-export-drops-inject.md)。

View File

@@ -25,7 +25,6 @@ import {
import type { ContentBlock } from '@deepseek-ai/dsh-llm'
import { SessionId } from '@deepseek-ai/dsh-session'
import type { SubagentResult, SubagentRun, SubagentStartRequest, SubagentStopReason } from '@deepseek-ai/dsh-subagent'
import { splitEnvChannels } from '@deepseek-ai/dsh-subprocess'
import type { SubprocessHandle, SubprocessSpawnSpec } from '@deepseek-ai/dsh-subprocess'
/** Fixed response to child permission requests: reject by default, or select the first allow option. */
@@ -51,8 +50,8 @@ export interface AcpRunSpec {
* parent env. A value here is forwarded even if its name matches the
* credential-scrub pattern (an explicit opt-in for the child's own creds).
* Explicit `DSH_*` entries are deployment-owned facts for the child harness
* (e.g. `DSH_PERMISSION_MODE`) and ride the seam's managed channel, which
* the scrubbed base reserves for current values.
* (e.g. `DSH_PERMISSION_MODE`); they simply merge after the scrub that
* dropped their stale ambient namesakes.
*/
env: Record<string, string>
/**
@@ -169,15 +168,14 @@ export async function startAcpRun(request: SubagentStartRequest, spec: AcpRunSpe
const id = SessionId(randomUUID())
// Keep diagnostics on parent stderr ('inherit'); only ACP output contributes
// to the result. The seam's scrub drops ambient credentials while spec.env
// (the child's own key) merges after it; explicit DSH_* entries ride the
// managed channel via the seam's split.
// to the result. The seam's scrub drops ambient credentials and DSH_* names
// while spec.env (the child's own key, its deployment facts) merges after it.
const child = spec.spawn({
argv: [spec.command, ...spec.args],
cwd: spec.cwd,
stdio: { stdin: 'pipe', stdout: 'pipe', stderr: 'inherit' },
graceMs: spec.disposeGraceMs,
...splitEnvChannels(spec.env),
env: spec.env,
})
/* v8 ignore start -- 'pipe' dispositions expose both streams by the seam contract; defensive. */
if (child.stdin === undefined || child.stdout === undefined) {

View File

@@ -120,11 +120,10 @@ describe('child env layering (through the subprocess seam)', () => {
}
})
it('routes explicit DSH_* config entries onto the managed channel', async () => {
it('forwards explicit DSH_* config entries to the child', async () => {
// A deployment sets child-harness facts like DSH_PERMISSION_MODE in
// config.env; the run must split them onto the seam's managed channel
// (the ordinary channel rejects the reserved namespace) and the child
// must still see the value.
// config.env; the seam's scrub drops only the AMBIENT namesakes, so the
// explicit entry merges after it and the child must see the value.
const ctx = await setup({ MOCK_ECHO_ENV: 'DSH_ACP_TEST_FACT', DSH_ACP_TEST_FACT: 'managed' })
const parent = { id: 'parent', session: { header: { cwd: process.cwd() } } } as unknown as Agent
const run = await ctx.subagents.start('acp', { prompt: [{ type: 'text' as const, text: 'p' }], parent, signal: new AbortController().signal })

View File

@@ -2,5 +2,5 @@
# side as of the last confirmed-consistent state. Both languages carry equal authority;
# after editing either side, bring the other along and re-record with:
# pnpm run verify-translation-pairing --write
README.md: 08cc2ce7d92569222b99992d0f4c43551a2c9623
README.zh.md: da230ba37d406a6ad4ec669ceead45f2c2dd7069
README.md: a62929e8ee1c1c852d99d9569f4118bc751e8150
README.zh.md: 99febdbe54988f54b74dbc2c80251ec1e9e671ef

View File

@@ -8,7 +8,7 @@ Local implementation of the [`@deepseek-ai/dsh-subprocess`](../subprocess/README
- **Detached process trees with platform-correct signalling** — POSIX children are spawned `detached` (own process group) and signalled by negative pgid with a direct-child fallback; Windows terminates the tree via `taskkill /PID <pid> /T /F` (injectable for tests). `terminate()` sends SIGTERM then SIGKILL after the spec's grace (OpenCode's escalation; pipelines and subshells die with the parent); `kill(signal)` sends exactly one signal and is a no-op after settlement; `dispose(graces)` runs stdin-EOF → SIGTERM → SIGKILL with caller-supplied windows and one memoized disposal per handle. After the leader exits, still-open pipes receive the same bounded drain grace so a surviving descendant cannot hold the outcome open indefinitely. ESRCH is tolerated; daemons that re-parent away from the group can still survive — the same caveat as the surveyed tools.
- **Per-stream dispositions** — `'pipe'` hands the raw stream to the caller untouched (protocol framing stays consumer-owned); `'inherit'` passes the parent descriptor through; collect mode keeps the in-memory TAIL beyond its cap (errors and results cluster at the end — pi/OpenCode rationale) while the FULL stream is appended to a private temp file when a spill cap is configured — omitting `spill` keeps only the tail, the diagnostic shape. A stream larger than the spill cap discards its now-incomplete spill and returns only the marked truncated tail; spill fds are sealed at settlement, and a failed final close withholds the path rather than advertising an incomplete file. Spill files are `0600` with random names under a lazily-created `0700` per-process directory.
- **Credential scrub + managed `DSH_*` merge** — `process.env` minus credential-shaped vars (`*KEY*`/`*SECRET*`/`*TOKEN*`) and all ambient `DSH_*` names; a spec's ordinary `env` merges after the scrub but rejects `DSH_*`; managed `dshEnv` rejects ordinary names and merges last, preventing stale nested-harness identity. Supplied stdin is written and closed; otherwise fd 0 is `/dev/null`. See the [stdin/env Agent Note](../../../.agents/notes/implemented/architecture/2026-06-30-bash-stdin-env-trusted-plugin-surface.md) and [managed environment Agent Note](../../../.agents/notes/implemented/feature/2026-07-10-agent-session-identity-and-log-location.md).
- **Credential scrub + explicit merge** — `process.env` minus credential-shaped vars (`*KEY*`/`*SECRET*`/`*TOKEN*`) and all ambient `DSH_*` names; the spec's explicit `env` merges after that scrub with no namespace validation, so a deliberately supplied credential or current `DSH_*` fact wins while stale nested-harness identity cannot leak in ambiently. Supplied stdin is written and closed; otherwise fd 0 is `/dev/null`. See the [stdin/env Agent Note](../../../.agents/notes/implemented/architecture/2026-06-30-bash-stdin-env-trusted-plugin-surface.md) and [managed environment Agent Note](../../../.agents/notes/implemented/feature/2026-07-10-agent-session-identity-and-log-location.md).
- **Offset-based reads** — collect-mode readers return deltas in whole-stream byte coordinates; the service never holds a cursor, so consumer-owned cursors (the bash background read path) and full-stream re-reads coexist, before and after settlement.
- **Terminate-and-join disposal** — the service retains live handles only so its own disposal can escalate every running tree and await its exit; settled and spawn-failed handles leave the live set on settlement.

View File

@@ -8,7 +8,7 @@
- **带平台正确信号发送的 detached 进程树**:POSIX 子进程使用 `detached` spawn(拥有独立进程组),信号以负 pgid 发送并以直接子进程作为回退;Windows 通过 `taskkill /PID <pid> /T /F` 终止进程树(可为测试注入)。`terminate()` 先发送 SIGTERM,经过 spec 的宽限期后再发送 SIGKILL(沿用 OpenCode 的升级策略;管道与子 shell 会随父进程一起结束);`kill(signal)` 恰好发送一个信号,结算后为空操作;`dispose(graces)` 以调用方提供的时间窗运行 stdin EOF→SIGTERM→SIGKILL 阶梯,dispose(资源释放)按句柄 memoize 化、只执行一次。组长进程退出后,仍然打开的管道也只获得同样有界的排空宽限期,因此存活的后代进程无法无限期地拖住结果不结算。系统会容忍 ESRCH;脱离该组重新挂载的 daemon 仍可能存活,这与调研工具的局限相同。
- **按流划分的处置方式**:`'pipe'` 把原始流原样交给调用方(协议分帧仍归消费方所有);`'inherit'` 直通父进程的描述符;收集模式(collect)在输出超过上限后于内存中保留尾部(错误与结果通常聚集在末尾,沿用 pi/OpenCode 的理由),并在配置了 spill 上限时把完整流追加到一个私有临时文件;省略 `spill` 则只保留尾部,即诊断尾部的形状。某条流大于 spill 上限时,会丢弃已不完整的 spill,仅返回带截断标记的尾部;spill 文件描述符在结算时封存,最终关闭失败时则不公布路径,以免声称存在不完整的文件。spill 文件权限为 `0600`、名称随机,位于按需延迟创建的 `0700` 每进程目录之下。
- **凭据清除 + 受管 `DSH_*` 合并**:以 `process.env` 为基础,移除形似凭据的变量(`*KEY*`/`*SECRET*`/`*TOKEN*`)和所有环境中已有的 `DSH_*` 名称;spec 的普通 `env` 在清除后合并,但会拒绝 `DSH_*`;受管 `dshEnv` 会拒绝普通名称并最后合并,防止陈旧的嵌套 harness 身份。提供的 stdin 会被写入后关闭;否则 fd 0 指向 `/dev/null`。参见 [stdin/env Agent Note(agent 决策记录)](../../../.agents/notes/implemented/architecture/2026-06-30-bash-stdin-env-trusted-plugin-surface.md)与[受管环境 Agent Note](../../../.agents/notes/implemented/feature/2026-07-10-agent-session-identity-and-log-location.md)。
- **凭据清除 + 显式合并**:以 `process.env` 为基础,移除形似凭据的变量(`*KEY*`/`*SECRET*`/`*TOKEN*`)和所有环境中已有的 `DSH_*` 名称;spec 的显式 `env` 在该清除之后合并且不做命名空间校验,因此有意提供的凭据或当前 `DSH_*` 事实会胜出,而陈旧的嵌套 harness 身份无法从环境中隐式漏入。提供的 stdin 会被写入后关闭;否则 fd 0 指向 `/dev/null`。参见 [stdin/env Agent Note(agent 决策记录)](../../../.agents/notes/implemented/architecture/2026-06-30-bash-stdin-env-trusted-plugin-surface.md)与[受管环境 Agent Note](../../../.agents/notes/implemented/feature/2026-07-10-agent-session-identity-and-log-location.md)。
- **基于偏移量的读取**:收集模式的读取器以全流字节坐标返回增量;服务自身从不持有游标,因此消费方自有的游标(bash 的后台读取路径)与完整流重读可以共存,结算前后皆然。
- **先终止再等待退出的 dispose**:服务保留存活句柄,只为让自身的 dispose 能对每个仍在运行的进程树执行升级并等待其退出;已结算与 spawn 失败的句柄在结算时即离开存活集合。

View File

@@ -16,10 +16,9 @@ import { tmpdir } from 'node:os'
import { join } from 'node:path'
import { setTimeout as sleepMs } from 'node:timers/promises'
import { deadline } from '@deepseek-ai/dsh-timeout'
import { DSH_ENV_PREFIX, scrubbedParentEnv } from '@deepseek-ai/dsh-subprocess'
import { scrubbedParentEnv } from '@deepseek-ai/dsh-subprocess'
import type {
CollectedOutput,
DshEnvironment,
SubprocessCollect,
SubprocessDisposeGraces,
SubprocessHandle,
@@ -29,28 +28,14 @@ import type {
} from '@deepseek-ai/dsh-subprocess'
/**
* Build a child environment from the scrubbed parent base, ordinary caller
* entries, and a managed `DSH_*` snapshot. Ordinary and managed entries
* reject the other channel's namespace before `dshEnv` merges last.
* @param extra - caller entries; `DSH_*` names are rejected.
* @param dshEnv - managed entries; non-`DSH_*` names are rejected.
* Build a child environment: explicit caller entries merge after the scrubbed
* parent base, so a deliberately supplied credential or current `DSH_*` fact
* wins over the scrub that dropped its ambient namesake.
* @param extra - explicit caller entries, merged verbatim after the scrub.
* @returns the environment to hand to `spawn` for the child process.
*/
export function childEnv(
extra?: Readonly<Record<string, string>>,
dshEnv?: DshEnvironment,
): NodeJS.ProcessEnv {
for (const key of Object.keys(extra ?? {})) {
if (key.startsWith(DSH_ENV_PREFIX)) {
throw new Error(`ordinary child env cannot set reserved variable "${key}"; use dshEnv`)
}
}
for (const key of Object.keys(dshEnv ?? {})) {
if (!key.startsWith(DSH_ENV_PREFIX)) {
throw new Error(`managed child env cannot set ordinary variable "${key}"; use env`)
}
}
return { ...scrubbedParentEnv(), ...extra, ...dshEnv }
export function childEnv(extra?: Readonly<Record<string, string>>): NodeJS.ProcessEnv {
return { ...scrubbedParentEnv(), ...extra }
}
/** Injectable knobs so tests can exercise spill and platform behavior deterministically. */
@@ -339,7 +324,7 @@ export function spawnSubprocess(spec: SubprocessSpawnSpec, internals: SpawnInter
const errMode = spec.stdio.stderr
const stdinMode = spec.stdio.stdin
const env = childEnv(spec.env, spec.dshEnv)
const env = childEnv(spec.env)
const child = spawn(program, args, {
cwd: spec.cwd,
env,

View File

@@ -2,7 +2,6 @@ import { mkdtempSync, readFileSync, statSync, unlinkSync } from 'node:fs'
import { tmpdir } from 'node:os'
import { dirname, join } from 'node:path'
import { describe, expect, it, vi } from 'vitest'
import type { DshEnvironment } from '@deepseek-ai/dsh-subprocess'
import { killGroup, OutputCollector, spawnSubprocess, taskkillProcessTree } from '../src/spawn.ts'
import type { SubprocessHandle, SubprocessOutputReader } from '@deepseek-ai/dsh-subprocess'
@@ -840,11 +839,13 @@ describe('environment and spill-file hardening', () => {
}
})
it('injects only the current trusted DSH environment after scrubbing ambient values', async () => {
it('forwards explicit DSH_* env entries while scrubbing ambient ones', async () => {
// Both facts through one explicit map: the ambient DSH_STALE is dropped by
// the scrub, and the deliberately supplied current values merge after it.
process.env.DSH_STALE = 'old-value'
try {
const result = await finish(spawnSubprocess(spec('echo "[${DSH_STALE:-absent}|$DSH_SHELL|$DSH_SESSION_ID]"', {
dshEnv: { DSH_SHELL: '1', DSH_SESSION_ID: 'current-session' },
env: { DSH_SHELL: '1', DSH_SESSION_ID: 'current-session' },
})))
expect(result.stdout.text.trim()).toBe('[absent|1|current-session]')
} finally {
@@ -852,17 +853,6 @@ describe('environment and spill-file hardening', () => {
}
})
it('rejects DSH variables on the ordinary env channel', () => {
expect(() => spawnSubprocess(spec('true', { env: { DSH_WRONG_CHANNEL: 'bad' } })))
.toThrow(/DSH_WRONG_CHANNEL.*dshEnv/)
})
it('rejects ordinary variables on the managed env channel', () => {
const invalid = { PATH: '/wrong-channel' } as unknown as DshEnvironment
expect(() => spawnSubprocess(spec('true', { dshEnv: invalid })))
.toThrow(/managed child env.*PATH.*use env/)
})
it('creates spill files with owner-only permissions and random names', async () => {
const result = await finish(spawnSubprocess(
spec('for i in $(seq 1 200); do printf "line-%04d\\n" $i; done', { stdoutMaxBytes: 500, stderrMaxBytes: 500 }),

View File

@@ -2,5 +2,5 @@
# side as of the last confirmed-consistent state. Both languages carry equal authority;
# after editing either side, bring the other along and re-record with:
# pnpm run verify-translation-pairing --write
README.md: 2cb7a5ebce404c440e625dea844ed28ceadb06f3
README.zh.md: a3211834e065359e813e8148a8f6a6a15f8f89b6
README.md: 65c2be2b77fbd7105208b86fa1b98a3c44013f94
README.zh.md: 98a0315914746a94350ceec53e4daff886054e64

View File

@@ -10,7 +10,7 @@ The subprocess seam (`ctx.subprocess`). The abstract `SubprocessService` exposes
- The spec is fully explicit — argv, cwd, per-stream stdio dispositions, grace — because deployment-varying defaults belong to the calling seam's config, not to a hidden subprocess-service default (the `dsh-bash` request/spec split is the owning template). `argv` is never shell-interpreted; a consumer that wants a shell passes `['bash', '-c', command]` itself.
- Stdio is Node-shaped per stream: `'pipe'` hands the caller the raw stream for its own protocol framing (LSP JSON-RPC, ACP ndjson), `'inherit'` passes the parent descriptor through for diagnostics, and collect mode (`{ maxBytes, spill? }`) buffers a bounded tail with an optional full-stream spill file. Collect readers take whole-stream byte offsets and never consume, so independent readers cannot steal one another's deltas; a read whose offset slid out of the in-memory tail is `lossy` and points at the spill file when one exists. Collected output stays readable after settlement.
- Termination is tree-scoped on every platform (POSIX detached groups with direct-child fallback; Windows `taskkill /T`): `kill(signal)` sends one signal Node-style and is a no-op after settlement, `terminate()` (and the spec's abort signal) escalates SIGTERM→grace→SIGKILL, `waitForExit()` observes the whole tree, and `dispose(graces)` runs the cooperative stdin-EOF→SIGTERM→SIGKILL ladder out-of-process children need — the manager reacts but never classifies why (callers own deadlines and cause classification).
- `scrubbedParentEnv()` / `SENSITIVE_ENV_PATTERN` are the one shared scrub definition: ambient credential-shaped and `DSH_*` names are dropped, explicit `env` merges after the scrub (a deliberately forwarded key survives), and `dshEnv` carries current harness facts on its own validated channel; `splitEnvChannels()` partitions a consumer config's single mixed env map onto those two channels (lsp-local servers and the ACP backend expose one map, and a configured `DSH_*` fact must ride the managed channel the ordinary one rejects). Spawners that cannot route through the service (node-pty backends, SDK-managed transports) import the scrub.
- `scrubbedParentEnv()` / `SENSITIVE_ENV_PATTERN` are the one shared scrub definition: ambient credential-shaped and `DSH_*` names are dropped, and the spec's explicit `env` merges after the scrub with no namespace validation — a deliberately forwarded credential or a current `DSH_*` fact survives precisely because it is an explicit caller opt-in, while the stale ambient namesake never reaches the child. Spawners that cannot route through the service (node-pty backends, SDK-managed transports) import the scrub.
- Disposal of the service terminates all still-running managed processes and awaits their exit.
See the [subprocess data-structure catalog](../../../docs/core-data-structures/subprocess.md) and the [seam Agent Note](../../../.agents/notes/implemented/architecture/2026-07-26-subprocess-seam.md).

View File

@@ -10,7 +10,7 @@
- spec 完全显式(argv、cwd、按流划分的 stdio 处置方式(disposition)、宽限期),因为随部署变化的默认值属于调用方 seam 的配置,而不属于某个隐藏的进程管理器默认值(`dsh-bash` 的 request/spec 拆分是这条规则的所属模板)。`argv` 绝不经过 shell 解释;需要 shell 的消费方自行传入 `['bash', '-c', command]`。
- stdio 按流采用 Node 形状:`'pipe'` 把原始流交给调用方做自己的协议分帧(LSP 的 JSON-RPC、ACP(Agent Client Protocol)的 ndjson),`'inherit'` 直通父进程描述符以承载诊断输出,收集模式(collect)`{ maxBytes, spill? }` 则缓冲一段有界尾部,外加可选的完整流 spill 文件。收集模式的读取器接受全流字节偏移量且从不消费,因此独立的读取器不会抢走彼此的增量;偏移量滑出内存尾部窗口的读取标记为 `lossy`,并在 spill 文件存在时指向它。收集到的输出在结算后仍可读取。
- 终止在每个平台上都以进程树为范围(POSIX 用 detached 进程组并以直接子进程回退;Windows 用 `taskkill /T`):`kill(signal)` 以 Node 风格只发送一个信号,结算后为空操作;`terminate()`(以及 spec 的 abort 信号)执行 SIGTERM→宽限期→SIGKILL 升级;`waitForExit()` 观察整棵进程树;`dispose(graces)` 运行进程外子进程所需的协作式 stdin EOF→SIGTERM→SIGKILL 阶梯。管理器只响应中止,但绝不判定原因(deadline 与原因分类归调用方所有)。
- `scrubbedParentEnv()` / `SENSITIVE_ENV_PATTERN` 是唯一一份共享的凭据清除定义:环境中形似凭据的名称与 `DSH_*` 名称都会被丢弃,显式 `env` 在清除之后合并(有意转发的键会保留下来),`dshEnv` 则经由自身带校验的通道携带当前 harness 事实;`splitEnvChannels()` 把消费方配置中单一的混合 env 映射按这两条通道切分(lsp-local 的服务器配置与 ACP 后端只暴露一个映射,而配置的 `DSH_*` 事实必须走受管通道,普通通道会拒绝它)。无法把 spawn 路由到该服务的调用点(node-pty 后端、由 SDK 管理的传输层)改为导入凭据清除函数。
- `scrubbedParentEnv()` / `SENSITIVE_ENV_PATTERN` 是唯一一份共享的凭据清除定义:环境中形似凭据的名称与 `DSH_*` 名称都会被丢弃,spec 的显式 `env` 在清除之后合并且不做命名空间校验——有意转发的凭据或当前 `DSH_*` 事实之所以能保留下来,正因为它是调用方的显式选择,而陈旧的同名环境值永远到不了子进程。无法把 spawn 路由到该服务的调用点(node-pty 后端、由 SDK 管理的传输层)改为导入凭据清除函数。
- 服务自身的 dispose(资源释放)会终止所有仍在运行的受管进程并等待其退出。
参见[进程管理器数据结构目录](../../../docs/core-data-structures/subprocess.md)与 [seam Agent Note(agent 决策记录)](../../../.agents/notes/implemented/architecture/2026-07-26-subprocess-seam.md)。

View File

@@ -12,7 +12,6 @@
import { Context, Service } from 'cordis'
import { DSH_ENV_PREFIX } from './types.ts'
import type { DshEnvironment, DshEnvironmentKey } from './types.ts'
import type { SubprocessHandle, SubprocessSpawnSpec } from './types.ts'
export { DSH_ENV_PREFIX } from './types.ts'
@@ -46,11 +45,10 @@ export const SENSITIVE_ENV_PATTERN = /KEY|SECRET|TOKEN/i
* The ambient parent environment minus credential-shaped names and minus all
* `DSH_*` names — the canonical base every harness child starts from. `PATH`,
* `HOME`, locale, and proxy variables survive, so child CLIs run normally;
* harness identity never leaks implicitly (a child that needs current `DSH_*`
* facts receives them through {@link SubprocessSpawnSpec.dshEnv}, and a
* deliberately forwarded credential goes through an explicit env layer, which
* merges after this scrub). Exported as a plain function so spawners that
* cannot route through the service (node-pty backends, SDK-managed
* harness identity never leaks implicitly (a deliberately forwarded
* credential or current `DSH_*` fact goes through the spec's explicit `env`,
* which merges after this scrub). Exported as a plain function so spawners
* that cannot route through the service (node-pty backends, SDK-managed
* transports) share the one scrub definition.
* @returns a fresh environment object safe to hand to a child spawn.
*/
@@ -62,27 +60,6 @@ export function scrubbedParentEnv(): Record<string, string> {
return env
}
/**
* Partition one mixed explicit-env map onto the spec's two channels: `DSH_*`
* names are deployment-owned facts for the child and take the managed
* {@link SubprocessSpawnSpec.dshEnv} channel (the ordinary channel rejects the
* reserved namespace), everything else stays ordinary `env`. For consumers
* whose configs expose a single env map (lsp-local servers, the ACP backend)
* rather than two channel-shaped fields.
* @param env - explicit entries from a consumer's config, both namespaces mixed.
* @returns the two spec channels, each safe for its validator.
*/
export function splitEnvChannels(env: Readonly<Record<string, string>>): { env: Record<string, string>; dshEnv: DshEnvironment } {
const ordinary: Record<string, string> = {}
const managed: Record<DshEnvironmentKey, string> = {}
const isDshKey = (key: string): key is DshEnvironmentKey => key.startsWith(DSH_ENV_PREFIX)
for (const [key, value] of Object.entries(env)) {
if (isDshKey(key)) managed[key] = value
else ordinary[key] = value
}
return { env: ordinary, dshEnv: managed }
}
declare module 'cordis' {
interface Context {
subprocess: SubprocessService

View File

@@ -93,19 +93,14 @@ export interface SubprocessSpawnSpec {
*/
signal?: AbortSignal | undefined
/**
* Ordinary environment entries merged onto the implementation's scrubbed
* parent base (see `scrubbedParentEnv`). `DSH_*` names are rejected and
* belong in {@link dshEnv}; a deliberately forwarded credential-shaped
* entry survives because this layer merges after the scrub.
* Explicit environment entries merged onto the implementation's scrubbed
* parent base (see `scrubbedParentEnv`), with no namespace validation:
* every entry is a deliberate caller opt-in, so a forwarded
* credential-shaped entry or a current `DSH_*` fact survives precisely
* because this layer merges after the scrub that drops its ambient
* namesake.
*/
env?: Record<string, string> | undefined
/**
* Harness-owned `DSH_*` variables for this execution. The scrubbed base has
* already discarded ambient `DSH_*` entries, so an unavailable current fact
* cannot inherit a stale value from the harness process; non-`DSH_*` names
* on this channel are rejected.
*/
dshEnv?: DshEnvironment | undefined
}
/**

View File

@@ -1,6 +1,6 @@
import { describe, expect, it } from 'vitest'
import { Context } from 'cordis'
import { scrubbedParentEnv, splitEnvChannels, SubprocessService } from '@deepseek-ai/dsh-subprocess'
import { scrubbedParentEnv, SubprocessService } from '@deepseek-ai/dsh-subprocess'
import type { SubprocessDisposeGraces, SubprocessHandle, SubprocessOutputRead, SubprocessSpawnSpec } from '@deepseek-ai/dsh-subprocess'
/**
@@ -72,10 +72,4 @@ describe('SubprocessService seam', () => {
delete process.env.SCRUB_PROBE_PLAIN
}
})
it('splitEnvChannels partitions a mixed map onto the two spec channels', () => {
const { env, dshEnv } = splitEnvChannels({ DSH_FACT: 'managed', PLAIN: 'ordinary', DEEPSEEK_API_KEY: 'explicit' })
expect(env).toEqual({ PLAIN: 'ordinary', DEEPSEEK_API_KEY: 'explicit' })
expect(dshEnv).toEqual({ DSH_FACT: 'managed' })
})
})