From 1b1ba96d4f608d7ea3bd9ea74d5b5ea1c169b14b Mon Sep 17 00:00:00 2001 From: imccyu <276526105+imccyu@users.noreply.github.com> Date: Thu, 9 Jul 2026 01:37:40 +0800 Subject: [PATCH] =?UTF-8?q?fix(tool-cordis):=20replace=20the=20pass-throug?= =?UTF-8?q?h=20ctx=20proxy=20with=20a=20whitelist=20fa=C3=A7ade?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Review finding (#220): the guarded proxy only special-cased ctx.tools, so mount code could reach an UNGUARDED context through ctx.root, ctx.extend(), or a service instance's .ctx, then ctx.root.tools.register({…}) to bypass the marker check and host-realm normalization — a raw vm-realm result would later error a real agent turn at the session-log plainness check. The sandbox ctx is now a whitelist façade, not a pass-through proxy: it exposes only what a mount needs — tools.register (marker-guarded), on/once, provide, the timer helpers, and injected services resolved through a guarded get — and denies every framework-plumbing member (root, parent, fiber, reflect, registry, extend, isolate, intercept, plugin, set, mixin, …) with a teaching error. Injected services are wrapped so a method returning a Context is rejected on the way back (the .ctx escape), closing the one indirect leak. There is no context-valued member left to reach; cross-mount provide/inject is untouched (the plugin's own inject and the fiber's pending/active gating are unchanged). ctx.plugin (child plugins) and ctx.set are denied by design; ctx.effect is deferred (FIXME). Adds tests/sandbox-context.spec.ts covering the escape class (root/extend/fiber/ plugin/set/… denied, the classic root.tools.register bypass, the .ctx escape, read-only writes) plus the async-service and symbol/in-operator paths for 100% coverage. RFC/README/tool-catalog/config-catalog updated; api-catalog.ts regenerated (also picks up the codeRuntime service that entered on the master merge and was left stale). --- docs/config-catalog.md | 2 +- ...6-07-08-self-referential-cordis-toolset.md | 8 +- docs/tool-catalog.md | 2 +- packages/cordis/tool-cordis/README.md | 2 +- .../cordis/tool-cordis/src/api-catalog.ts | 31 ++++ packages/cordis/tool-cordis/src/guard.ts | 169 ++++++++++++++---- packages/cordis/tool-cordis/src/index.ts | 19 +- .../cordis/tool-cordis/tests/mount.spec.ts | 4 +- .../tool-cordis/tests/sandbox-context.spec.ts | 156 ++++++++++++++++ 9 files changed, 348 insertions(+), 45 deletions(-) create mode 100644 packages/cordis/tool-cordis/tests/sandbox-context.spec.ts diff --git a/docs/config-catalog.md b/docs/config-catalog.md index d466bff4c0..2e6be3a1ea 100644 --- a/docs/config-catalog.md +++ b/docs/config-catalog.md @@ -686,7 +686,7 @@ export interface Config { } ``` -Source: [`packages/cordis/tool-cordis/src/index.ts:49`](../packages/cordis/tool-cordis/src/index.ts) +Source: [`packages/cordis/tool-cordis/src/index.ts:53`](../packages/cordis/tool-cordis/src/index.ts) ## `@deepseek-ai/dsh-tool-fs` diff --git a/docs/rfc/implemented/feature/2026-07-08-self-referential-cordis-toolset.md b/docs/rfc/implemented/feature/2026-07-08-self-referential-cordis-toolset.md index 26daf4a844..2ba398301b 100644 --- a/docs/rfc/implemented/feature/2026-07-08-self-referential-cordis-toolset.md +++ b/docs/rfc/implemented/feature/2026-07-08-self-referential-cordis-toolset.md @@ -12,7 +12,7 @@ First, model-written registration must be validated where it happens: a malforme The toolset ships as [`@deepseek-ai/dsh-tool-cordis`](../../../../packages/cordis/tool-cordis/README.md) — a new top-level `packages/cordis/` group — and is demoed by [`examples/cordis-agent`](../../../../examples/cordis-agent/README.md). It gives the model three tools over the live cordis runtime it is running inside: inspect it, mount model-written plugins into it, dispose them again. -The trust stance, stated once and threaded through the rest: the `node:vm` sandbox isolates the global context only — it prevents accidental global pollution, not malice. The `ctx` handed to a mounted plugin's `apply` is the real, fully privileged runtime handle; handing the model that handle is the point of the toolset. A deployment loads this plugin exactly as deliberately as it grants a bash tool — an opt-in capability in the app's `cordis.yml`, never a product default. +The trust stance, stated once and threaded through the rest: the `node:vm` sandbox isolates the global context only — it prevents accidental global pollution, not malice — and the `ctx` a mounted plugin's `apply` receives is a whitelist façade that narrows the *surface* (framework internals withheld) but not the *privilege* of what it exposes. The verbs the façade does expose reach the real runtime: a mounted tool can shell out through `ctx.bash`, read the filesystem through `ctx.fs`, reach the network through `ctx.web`. Neither the sandbox nor the façade is a security boundary; handing the model this power is the point of the toolset. A deployment loads this plugin exactly as deliberately as it grants a bash tool — an opt-in capability in the app's `cordis.yml`, never a product default. ### The three tools @@ -30,7 +30,7 @@ Mount code runs via `vm.createContext` + `runInContext`, wrapped as the body of Sandbox globals are deliberately small: a tagged write-through `console` (`[cordis:] …` on the host stdout/stderr, so a listener that fires long after the mount call still lands somewhere the user sees), the `harness.defineTool` / `harness.registerTool` registration pair, the encoding primitives fresh vm contexts lack (`btoa`/`atob` as host closures over `Buffer` — a sanctioned exception, `Buffer` itself is never exposed — plus `TextEncoder`/`TextDecoder`), and callable traps over the withheld Node APIs (`require`, `setTimeout`/`setInterval`/`setImmediate`/`clearTimeout`/`clearInterval`, `fetch`) that throw a redirect naming the cordis alternative. Only function-shaped globals are trapped; `process` and `Buffer` stay `undefined` so a `typeof` feature probe stays inert rather than detonating a throwing accessor. -Three boundary mechanisms make model-written code behave correctly across the realm seam. **Dual-realm `instanceof`**: most objects sandbox code touches are host-realm (tool `args`, event payloads, service returns), so a plain `x instanceof Array` in the vm would silently be false — a per-sandbox prelude gives the vm realm's own constructors a `Symbol.hasInstance` that checks both the vm constructor and its host counterpart, patching only vm-realm globals. **Realm normalization of tool results**: objects built inside the vm carry the vm realm's `Object.prototype`, which the session log's append-time plainness check (`isJsonValue` in `dsh-session`, a prototype-identity comparison) rejects, so the sandbox's `harness.defineTool` JSON round-trips every `execute` return into the host realm — which also projects it onto exactly what the log durably stores. **Guarded registration**: the `ctx` a mounted plugin receives is a proxy whose `tools.register` accepts only definitions returned by `harness.defineTool` (a marker symbol), so every dynamic tool passes SchemaSpec validation and realm normalization; everything else on `ctx` passes through with correct `this` binding, which is what keeps cross-mount `provide`/`inject` working. +Three boundary mechanisms make model-written code behave correctly across the realm seam. **Dual-realm `instanceof`**: most objects sandbox code touches are host-realm (tool `args`, event payloads, service returns), so a plain `x instanceof Array` in the vm would silently be false — a per-sandbox prelude gives the vm realm's own constructors a `Symbol.hasInstance` that checks both the vm constructor and its host counterpart, patching only vm-realm globals. **Realm normalization of tool results**: objects built inside the vm carry the vm realm's `Object.prototype`, which the session log's append-time plainness check (`isJsonValue` in `dsh-session`, a prototype-identity comparison) rejects, so the sandbox's `harness.defineTool` JSON round-trips every `execute` return into the host realm — which also projects it onto exactly what the log durably stores. **A whitelist context façade**: the `ctx` a mounted plugin's `apply` receives is NOT the real context nor a pass-through proxy over it — it is a façade exposing only what a mount legitimately needs (`tools.register` marker-guarded, `on`/`once`, `provide`, the timer helpers, and injected services resolved through a guarded `get`), with every framework-plumbing member (`root`, `parent`, `fiber`, `reflect`, `registry`, `extend`, `isolate`, `intercept`, `plugin`, `set`, `mixin`, …) denied with a teaching error. This closes an escape *class* rather than a single hole: a proxy that merely special-cased `ctx.tools` still handed back the raw context through `ctx.root`, `ctx.extend()`, or a service instance's `.ctx`, and mount code could then `ctx.root.tools.register({…})` to bypass the marker check and realm normalization — a raw vm-realm result then errors a real agent turn at the plainness check. The façade has no context-valued member to reach, and the one indirect leak (an injected-service method returning a `Context`) is rejected on the way back to sandbox code; cross-mount `provide`/`inject` keeps working because the plugin's own `inject` and the fiber's pending/active gating are untouched — only the `apply`-time `ctx` surface is narrowed. Boundary errors are written around the mistakes models actually make (see [Consequences](#consequences) for how each was found), and the boundary normalizes rather than lectures wherever the input has exactly one meaning: schema `parameters` accept the JSON-Schema dialect models write by strong prior — the `{ type: 'object', properties, required: […] }` wrapper unwraps to the SchemaSpec DSL (the `required` array becoming per-property flags, at any nesting level), `type: 'integer'` maps to `number`, and `required: false` reads as optional — while genuinely meaningless input is rejected with the vocabulary enumerated (an unknown type lists the five valid ones; a non-boolean `required` names the rule). The remaining teaching errors: an unbalanced `});` closing gets the vm's offending source line plus a "code is a function body" reminder; TypeScript syntax gets the remove-annotations fix (detected on the failing line only, so an ` as ` inside a description string does not misfire); a forgotten `return` gets the two valid plugin forms; a Node built-in call gets the redirect to its cordis service; a tool-name collision on re-mount gets the unmount-first-then-remount recipe. @@ -40,7 +40,7 @@ Every dynamic mount is a child of a single `cordis-dynamic` group fiber, itself ### Cross-mount composition via provide/inject -Mounts relate to each other through ordinary cordis service semantics, with their ids as the lifecycle handles: mount A calls `ctx.provide('foo', value)`, mount B declares `inject: ['foo']` and activates the moment `foo` exists; mounted first, B stays pending and names the missing service; unmounting A sends B back to pending (its registrations unwound) and a later re-provide re-runs B's `apply` through the same guarded context; a duplicate provide fails loud with the owning fiber named. One realm caveat: a service value provided by a mount is a vm-realm object — method calls on it work from anywhere, but consumers must not assume host prototypes on it. +Mounts relate to each other through ordinary cordis service semantics, with their ids as the lifecycle handles: mount A calls `ctx.provide('foo', value)`, mount B declares `inject: ['foo']` and activates the moment `foo` exists; mounted first, B stays pending and names the missing service; unmounting A sends B back to pending (its registrations unwound) and a later re-provide re-runs B's `apply` through a fresh sandbox façade; a duplicate provide fails loud with the owning fiber named. One realm caveat: a service value provided by a mount is a vm-realm object — method calls on it work from anywhere, but consumers must not assume host prototypes on it. ### The generated API catalog @@ -73,7 +73,7 @@ The correctness investment therefore goes where it pays for every capability at **A new `cordis/mount` session event.** A durable provenance event recording each mount (source, name) has clear precedent (`hook/invoked`, `compact/start`). It was declined for v1: mount and unmount are already visible as `tool/call` / `tool/result` pairs and the tool-set change is already logged as a request-header delta, so a dedicated event would only duplicate the record. It remains addable if an audit use case needs mount provenance separable from the tool call. -**A hardened / capability-restricted sandbox.** Trapping Node built-ins might suggest an intent to sandbox for safety. It is explicitly not that: the traps redirect the model toward cordis services (and away from leak-prone Node timers) for correctness and inspectability, but `ctx` is fully privileged and the vm is not a security boundary. A real security boundary (separate process, permission prompts) was out of scope for a dev/opt-in toolset and would fight the entire point — handing the model the live runtime. +**A hardened / capability-restricted sandbox.** Trapping Node built-ins and handing mount code a whitelist façade rather than the raw context might suggest an intent to sandbox for safety. It is explicitly not that: the traps and the façade narrow the *surface* mount code sees — steering it onto cordis services and away from leak-prone Node built-ins and framework internals — for correctness and to close the unguarded-context escape, but the capabilities the façade exposes (`ctx.bash`, `ctx.fs`, `ctx.web`) reach the real runtime, so it is not a security boundary. A real one (separate process, permission prompts) was out of scope for a dev/opt-in toolset and would fight the entire point — handing the model the live runtime. ## Consequences diff --git a/docs/tool-catalog.md b/docs/tool-catalog.md index ec1546592c..990e135b1c 100644 --- a/docs/tool-catalog.md +++ b/docs/tool-catalog.md @@ -136,7 +136,7 @@ Source: [`packages/cordis/tool-cordis/src/index.ts`](../packages/cordis/tool-cor ### `cordis_mount` -Mount a NEW cordis plugin into the live runtime that is running THIS agent (self-modification). `code` runs as the body of an async JavaScript function in an isolated sandbox and MUST `return` a plugin. Two forms: FUNCTION form `return (ctx) => { … }` — cannot declare inject, uses whatever services are on the parent context, and accessing a service without inject (e.g. ctx.bash) throws; use it only when you need no injected services. OBJECT form `return { name?, inject: ['bash', 'llm', …], apply(ctx) { … } }` — declares dependencies, and cordis activates the plugin only after the services exist; PREFER this form for any plugin that needs bash, llm, sessions, etc. BEFORE calling a service from your code, read cordis_inspect what:"api" — it lists method signatures AND the type shapes of their arguments/returns (do not guess a field's type; e.g. a bash run's stdout is an object, not a string). Inside `apply`, use the standard cordis API: `ctx.on(event, listener)` to observe events (see cordis_inspect what:"events"), or call `harness.registerTool(ctx, harness.defineTool({ name, description, parameters: { text: { type: 'string', required: true } }, async execute(args) { … } }))` to give yourself a new tool — it becomes callable on your NEXT step. Tool parameters: each key IS a property — { type: 'string'|'number'|'boolean'|'object'|'array', required?: true, description?, enum?, items?, properties? }; a JSON-Schema-style { type: 'object', properties, required: […] } wrapper and type 'integer' are also accepted and normalized. A tool's `execute` MUST return an ARRAY of content blocks, e.g. `return [{ type: 'text', text: someString }]` — never a bare string. Mounts can COMPOSE: one plugin may `ctx.provide('name', value)` a service and another may declare `inject: ['name']` to consume it — the consumer stays pending until the provider exists and returns to pending when the provider is unmounted. Everything registered inside `apply` is cleaned up automatically on unmount. Sandbox globals: `console` (tagged `[cordis:]`, writes through to the harness terminal), `harness.defineTool`, `harness.registerTool`, `btoa`, `atob`, `TextEncoder`, `TextDecoder`. Node APIs are DISABLED — do filesystem/network/timer work through the cordis services, never Node built-ins: `require`, `setTimeout`/`setInterval`, and `fetch` throw redirect errors; `process` and `Buffer` are undefined. Instead use inject: ['fs'] + ctx.fs for files, inject: ['web'] + ctx.web for HTTP, inject: ['bash'] + ctx.bash for processes, and inject: ['timer'] + ctx.setTimeout/ctx.setInterval for timing (fiber effects, auto-cleaned on unmount) — cordis_inspect what:"api" shows what THIS runtime provides. Write PLAIN JavaScript, not TypeScript (no `as`, no type annotations). Cautions: (1) waterfall events (e.g. tools/pre-execute) hand the listener a trailing `next` callback which MUST be called — returning without `next()` VETOES the call; prefer plain notification events unless you intend to intercept. (2) Never await something that only resolves after the current turn (your code runs INSIDE a tool call of that turn — it would deadlock). (3) The sandbox prevents accidental global pollution, not malice: `ctx` is the real, fully privileged runtime handle. +Mount a NEW cordis plugin into the live runtime that is running THIS agent (self-modification). `code` runs as the body of an async JavaScript function in an isolated sandbox and MUST `return` a plugin. Two forms: FUNCTION form `return (ctx) => { … }` — cannot declare inject, uses whatever services are on the parent context, and accessing a service without inject (e.g. ctx.bash) throws; use it only when you need no injected services. OBJECT form `return { name?, inject: ['bash', 'llm', …], apply(ctx) { … } }` — declares dependencies, and cordis activates the plugin only after the services exist; PREFER this form for any plugin that needs bash, llm, sessions, etc. BEFORE calling a service from your code, read cordis_inspect what:"api" — it lists method signatures AND the type shapes of their arguments/returns (do not guess a field's type; e.g. a bash run's stdout is an object, not a string). Inside `apply`, use the standard cordis API: `ctx.on(event, listener)` to observe events (see cordis_inspect what:"events"), or call `harness.registerTool(ctx, harness.defineTool({ name, description, parameters: { text: { type: 'string', required: true } }, async execute(args) { … } }))` to give yourself a new tool — it becomes callable on your NEXT step. Tool parameters: each key IS a property — { type: 'string'|'number'|'boolean'|'object'|'array', required?: true, description?, enum?, items?, properties? }; a JSON-Schema-style { type: 'object', properties, required: […] } wrapper and type 'integer' are also accepted and normalized. A tool's `execute` MUST return an ARRAY of content blocks, e.g. `return [{ type: 'text', text: someString }]` — never a bare string. Mounts can COMPOSE: one plugin may `ctx.provide('name', value)` a service and another may declare `inject: ['name']` to consume it — the consumer stays pending until the provider exists and returns to pending when the provider is unmounted. Everything registered inside `apply` is cleaned up automatically on unmount. Sandbox globals: `console` (tagged `[cordis:]`, writes through to the harness terminal), `harness.defineTool`, `harness.registerTool`, `btoa`, `atob`, `TextEncoder`, `TextDecoder`. Node APIs are DISABLED — do filesystem/network/timer work through the cordis services, never Node built-ins: `require`, `setTimeout`/`setInterval`, and `fetch` throw redirect errors; `process` and `Buffer` are undefined. Instead use inject: ['fs'] + ctx.fs for files, inject: ['web'] + ctx.web for HTTP, inject: ['bash'] + ctx.bash for processes, and inject: ['timer'] + ctx.setTimeout/ctx.setInterval for timing (fiber effects, auto-cleaned on unmount) — cordis_inspect what:"api" shows what THIS runtime provides. Write PLAIN JavaScript, not TypeScript (no `as`, no type annotations). Cautions: (1) waterfall events (e.g. tools/pre-execute) hand the listener a trailing `next` callback which MUST be called — returning without `next()` VETOES the call; prefer plain notification events unless you intend to intercept. (2) Never await something that only resolves after the current turn (your code runs INSIDE a tool call of that turn — it would deadlock). (3) Your `ctx` is a restricted façade: you can register tools, observe events, provide/consume services, and use timers, but framework internals (ctx.root, ctx.fiber, ctx.extend, ctx.plugin, …) are withheld. It is not a security boundary though — the services you inject (e.g. ctx.bash) reach the real runtime. ```json { diff --git a/packages/cordis/tool-cordis/README.md b/packages/cordis/tool-cordis/README.md index b369a7209b..651f3273d9 100644 --- a/packages/cordis/tool-cordis/README.md +++ b/packages/cordis/tool-cordis/README.md @@ -12,7 +12,7 @@ Exact model-facing schemas: [the generated tool catalog](../../../docs/tool-cata ## Trust stance -The sandbox isolates the global context only — it is not a security boundary. No Node API is provided: `require`, the timers, and `fetch` are callable traps that throw a redirect to the cordis alternative (`ctx.fs` / `ctx.web` / `ctx.bash` / `inject: ['timer']` + `ctx.setTimeout`); `process` and `Buffer` are `undefined`; `globalThis` writes stay inside. The `ctx` a mounted plugin's `apply` receives is the real, fully privileged runtime handle; load this plugin as deliberately as you would grant a bash tool. +The sandbox isolates the global context only — it is not a security boundary. No Node API is provided: `require`, the timers, and `fetch` are callable traps that throw a redirect to the cordis alternative (`ctx.fs` / `ctx.web` / `ctx.bash` / `inject: ['timer']` + `ctx.setTimeout`); `process` and `Buffer` are `undefined`; `globalThis` writes stay inside. The `ctx` a mounted plugin's `apply` receives is a whitelist façade — register tools, observe events, provide/consume services, use timers; framework internals (`ctx.root`, `ctx.fiber`, `ctx.extend`, `ctx.plugin`, …) are withheld — but the capabilities it does expose reach the real runtime, so load this plugin as deliberately as you would grant a bash tool. ## Config diff --git a/packages/cordis/tool-cordis/src/api-catalog.ts b/packages/cordis/tool-cordis/src/api-catalog.ts index 39f1a12f47..83170d8dc1 100644 --- a/packages/cordis/tool-cordis/src/api-catalog.ts +++ b/packages/cordis/tool-cordis/src/api-catalog.ts @@ -88,6 +88,13 @@ export const SERVICE_API: readonly ServiceApiEntry[] = [ 'onTaskDone(listener: BashTaskListener): () => void', ], }, + { + key: 'codeRuntime', + summary: 'Abstract code-execution service.', + methods: [ + 'abstract run(request: CodeRunRequest): Promise', + ], + }, { key: 'compact', summary: 'Abstract compaction service.', @@ -429,6 +436,30 @@ export const TYPE_API: readonly TypeApiEntry[] = [ name: 'CallId', declaration: 'export type CallId = Branded<\'CallId\'>;', }, + { + name: 'CodeBindingFunction', + declaration: 'export type CodeBindingFunction = (args: unknown) => Promise;', + }, + { + name: 'CodeBindingNamespace', + declaration: 'export interface CodeBindingNamespace {\n global: string;\n functions: Record;\n}', + }, + { + name: 'CodeLogEntry', + declaration: 'export interface CodeLogEntry {\n source: \'console\' | \'stdout\' | \'stderr\';\n level?: \'log\' | \'info\' | \'warn\' | \'error\' | \'debug\';\n text: string;\n}', + }, + { + name: 'CodeRunFailure', + declaration: 'export interface CodeRunFailure {\n kind: \'exception\' | \'timeout\' | \'abort\' | \'worker-exit\';\n message: string;\n}', + }, + { + name: 'CodeRunRequest', + declaration: 'export interface CodeRunRequest {\n program: string;\n bindings: CodeBindingNamespace[];\n signal?: AbortSignal;\n}', + }, + { + name: 'CodeRunResult', + declaration: 'export interface CodeRunResult {\n value?: unknown;\n logs: CodeLogEntry[];\n error?: CodeRunFailure;\n}', + }, { name: 'CollectedOutput', declaration: 'export interface CollectedOutput {\n text: string;\n truncated: boolean;\n spillPath?: string;\n}', diff --git a/packages/cordis/tool-cordis/src/guard.ts b/packages/cordis/tool-cordis/src/guard.ts index 998997a8f9..07cd0f1df3 100644 --- a/packages/cordis/tool-cordis/src/guard.ts +++ b/packages/cordis/tool-cordis/src/guard.ts @@ -2,18 +2,35 @@ * The registration boundary between sandboxed mount code and the real runtime: * SchemaSpec normalization + validation with teaching errors, the * marker-guarded `harness.defineTool` / `harness.registerTool` pair, the - * guarded `ctx` proxy a mounted plugin receives, and the plugin-shape helpers - * the mount lifecycle narrows sandbox return values with. + * SANDBOX CONTEXT FAÇADE a mounted plugin's `apply` receives in place of the + * real `ctx`, and the plugin-shape helpers the mount lifecycle narrows sandbox + * return values with. * - * Two realm facts drive the design. Objects built inside the vm carry the vm + * The façade is a WHITELIST, not a pass-through proxy. Mount code needs to do + * exactly four things — register a tool, listen to an event, provide a service, + * call an injected service (timers included) — so the façade exposes only those + * verbs and the injected services, each individually wrapped. Every framework + * plumbing member (`root`, `parent`, `scope`, `fiber`, `reflect`, `registry`, + * `events`, `extend`, `isolate`, `intercept`, `plugin`, `set`, `mixin`, …) is + * DENIED with a teaching error rather than passed through. This closes an + * entire escape class at once: a pass-through proxy that only special-cased + * `ctx.tools` still handed back the raw context through `ctx.root`, + * `ctx.extend()`, or a service instance's `.ctx`, and mount code could then + * `ctx.root.tools.register({…})` to bypass the marker check and host-realm + * normalization — a raw vm-realm result then errors a real agent turn at the + * session-log plainness check. The whitelist has no such hole: there is no + * context-valued member to reach, and any injected-service method that returns + * a `Context` is rejected (harness services never do — see {@link denyContext}). + * + * Two realm facts drive the tool path. Objects built inside the vm carry the vm * realm's `Object.prototype`, and the session log's append-time plainness check * (`dsh-session`'s `isJsonValue`, a prototype-identity comparison) rejects * foreign-realm data — so every dynamic tool's `execute` return is JSON * round-tripped into the host realm before it reaches the registry, and the * schema itself is rebuilt as fresh host-realm objects. And a malformed tool * schema must fail at REGISTRATION, not when a later request assembles it — so - * dynamic `ctx.tools.register` calls accept only definitions produced by the - * sandbox's `harness.defineTool`, which normalizes `parameters` up front. + * dynamic tool registration accepts only definitions produced by the sandbox's + * `harness.defineTool`, which normalizes `parameters` up front. * * Normalize, don't lecture, where the input has exactly one meaning: models * write the JSON-Schema dialect by strong prior (the `{ type: 'object', @@ -26,7 +43,8 @@ * @module @deepseek-ai/dsh-tool-cordis/guard */ -import type { Context, Plugin } from 'cordis' +import { Context } from 'cordis' +import type { Plugin } from 'cordis' import { defineTool } from '@deepseek-ai/dsh-tools' import type { ToolDefinition, ToolExecuteReturn } from '@deepseek-ai/dsh-tools' @@ -153,31 +171,116 @@ export function sandboxRegisterTool(ctx: Context, tool: unknown): () => void { return ctx.tools.register(tool) } -function bindMethod(value: unknown, target: object): unknown { - if (typeof value !== 'function') return value - return (...args: unknown[]): unknown => Reflect.apply(value, target, args) as unknown +/** + * The verbs a mounted plugin may reach through the sandbox `ctx` façade, + * beyond its injected services. `on`/`once` observe events, `provide` exposes + * a service to other mounts, and the timer helpers schedule work — each a + * fiber effect that unwinds on unmount. Everything else on a real cordis `ctx` + * is framework plumbing and is denied. Forwarded LAZILY: the timer helpers are + * mixin accessors that throw `without inject` when read on a plugin that did + * not inject `timer`, so the façade reads `ctx[verb]` only at call time — the + * plugin that never touches a timer never trips that, and one that does gets + * cordis's own inject error at the call site. + */ +const CTX_VERBS = new Set(['on', 'once', 'provide', 'timeout', 'interval', 'setTimeout', 'setInterval', 'throttle', 'debounce']) + +/** + * The tool-registry façade: only `register` (marker-guarded), plus the + * read-only `schemas` / `get` a mount may legitimately want. No other registry + * method (nothing that could re-enter the raw context) is exposed. + */ +function sandboxTools(ctx: Context): Record { + return { + register: (tool: unknown): (() => void) => sandboxRegisterTool(ctx, tool), + schemas: () => ctx.tools.schemas(), + get: (name: string) => ctx.tools.get(name), + } } -function guardedContext(ctx: Context): Context { - const tools = new Proxy(ctx.tools, { +/** + * Reject any injected-service return that is a cordis `Context`. Harness + * services return data, never a context; a value that is one would be a + * fresh, unguarded handle back into the runtime — the exact escape the façade + * exists to close — so it fails loud instead of reaching sandbox code. + */ +function denyContext(value: unknown, service: string): unknown { + if (value instanceof Context) { + throw new Error( + `service "${service}" returned a cordis Context, which the sandbox does not expose. ` + + 'Operate through your own plugin ctx (ctx.on / ctx.provide / ctx.tools.register) ' + + 'and the services you inject — never another context.', + ) + } + return value +} + +/** + * Wrap an injected service so its methods forward to the real instance but + * their return values pass through {@link denyContext}. Non-function members + * (plain data) pass through as-is; a returned Promise is guarded on resolve. + */ +function guardedService(service: object, name: string): unknown { + return new Proxy(service, { get(target, prop) { - if (prop === 'register') { - return (tool: unknown): () => void => sandboxRegisterTool(ctx, tool) - } const value = Reflect.get(target, prop, target) as unknown - return bindMethod(value, target) + if (typeof value !== 'function') return denyContext(value, name) + return (...args: unknown[]): unknown => { + const result = Reflect.apply(value, target, args) as unknown + if (result instanceof Promise) return result.then(v => denyContext(v, name)) + return denyContext(result, name) + } }, }) - return new Proxy(ctx, { - get(target, prop) { +} + +/** + * The sandbox context façade handed to a mounted plugin's `apply` in place of + * the real `ctx`. A whitelist (see the module doc): the registration/eventing + * verbs, the timer helpers, a guarded `tools`, and injected services resolved + * through a guarded `get` / property access. Every framework-plumbing member + * is denied with a teaching error; there is no context-valued member to reach. + */ +function sandboxContext(ctx: Context): Context { + const tools = sandboxTools(ctx) + // Resolve a named service to a guarded wrapper, or undefined when absent. + const resolveService = (name: string): unknown => { + if (name === 'tools') return tools + const service: unknown = ctx.get(name) + return service === undefined ? undefined : guardedService(service as object, name) + } + const get = (name: string): unknown => resolveService(name) + return new Proxy({}, { + get(_target, prop) { if (prop === 'tools') return tools - if (prop === 'get') { - return (service: string): unknown => service === 'tools' ? tools : target.get(service) + if (prop === 'get') return get + if (typeof prop !== 'string') return undefined + // Lazy verb forwarder — reads `ctx[verb]` only when called, so a plugin + // that never uses a timer never triggers the timer mixin's inject check. + if (CTX_VERBS.has(prop)) { + return (...args: unknown[]): unknown => { + const method = ctx[prop as keyof Context] + return Reflect.apply(method as (...a: unknown[]) => unknown, ctx, args) + } } - const value = Reflect.get(target, prop, target) as unknown - return bindMethod(value, target) + // A declared-and-injected service reads as a ctx property; resolve it + // through the same guard. Absent → the deny path (framework plumbing, + // an un-injected service, or a typo) with one teaching error. + const service = resolveService(prop) + if (service !== undefined) return service + throw new Error( + `sandbox ctx does not expose "${prop}". Available: ctx.tools.register / ctx.on / ctx.provide / ` + + 'the timer helpers (ctx.setTimeout, ctx.interval, …) and any service you declared in inject. ' + + 'Framework internals (root, fiber, registry, extend, plugin, …) are withheld by design.', + ) }, - }) + // A façade is not the real ctx; block writes rather than let mount code + // stash state on a throwaway object and think it persisted. + set(_target, prop) { + throw new Error(`sandbox ctx is read-only; cannot assign "${String(prop)}"`) + }, + has: (_target, prop) => prop === 'tools' || prop === 'get' + || (typeof prop === 'string' && (CTX_VERBS.has(prop) || resolveService(prop) !== undefined)), + }) as unknown as Context } /** @@ -194,13 +297,19 @@ export function isPlugin(value: unknown): value is Plugin { } /** - * Wrap a plugin so its `apply` receives a guarded context (`tools.register` - * only accepts tools from `harness.defineTool`). Both function-form and - * object-form plugins go through the same guard; everything else on the - * context — `on`, `provide`, `inject` resolution — passes through with correct - * `this` binding, so cross-mount provide/inject works unmodified. + * Wrap a plugin so its `apply` receives the sandbox context façade instead of + * the real `ctx` (see {@link sandboxContext} and the module doc). Both + * function-form and object-form plugins go through the same wrap; the plugin's + * own `inject` declaration is preserved (cordis reads it from the plugin + * object, and pending/active gating happens on the real fiber before `apply` + * runs), so cross-mount provide/inject works unmodified. + * + * `ctx.effect(customCleanup)` is deliberately absent from the façade for now — + * `on` / `provide` / `tools.register` cover every mount seen so far, and each + * is already a fiber effect. FIXME(sandbox-effect): expose a guarded `effect` + * once a real mount needs a bespoke disposer. * @param plugin - the plugin the mount code returned. - * @returns an equivalent plugin whose `apply` sees the guarded context. + * @returns an equivalent plugin whose `apply` sees the sandbox context façade. */ export function guardedPlugin(plugin: Plugin): Plugin { if (typeof plugin === 'function') { @@ -208,7 +317,7 @@ export function guardedPlugin(plugin: Plugin): Plugin { return { name: pluginName(plugin), apply(ctx: Context, config?: unknown) { - return functionPlugin(guardedContext(ctx), config) + return functionPlugin(sandboxContext(ctx), config) }, } } @@ -216,7 +325,7 @@ export function guardedPlugin(plugin: Plugin): Plugin { return { ...plugin, apply(ctx: Context, config?: unknown) { - return objectPlugin.apply(guardedContext(ctx), config) + return objectPlugin.apply(sandboxContext(ctx), config) }, } } diff --git a/packages/cordis/tool-cordis/src/index.ts b/packages/cordis/tool-cordis/src/index.ts index 875637bfc4..f7d1958625 100644 --- a/packages/cordis/tool-cordis/src/index.ts +++ b/packages/cordis/tool-cordis/src/index.ts @@ -18,10 +18,14 @@ * this plugin. Design home: * docs/rfc/implemented/feature/2026-07-08-self-referential-cordis-toolset.md. * - * The vm sandbox guards against ACCIDENTAL global pollution only — it is not a - * security boundary. The `ctx` handed to the mounted plugin's `apply` is the - * real, fully privileged runtime handle; that is the point of the toolset, so - * a deployment loads this plugin as deliberately as it grants a bash tool. + * The vm sandbox guards against ACCIDENTAL global pollution only, and the `ctx` + * a mounted plugin's `apply` receives is a WHITELIST façade (register a tool, + * observe events, provide/consume services, use timers — framework internals + * withheld; see the guard module). Neither is a security boundary: the verbs + * the façade DOES expose reach the real runtime unsandboxed (a mounted tool can + * shell out through `ctx.bash`), so a deployment loads this plugin as + * deliberately as it grants a bash tool. Design home: + * docs/rfc/implemented/feature/2026-07-08-self-referential-cordis-toolset.md. * * Plugin export shape: named exports, NO default. The cordis Loader's * `unwrapExports` does `exports.default ?? exports`, so a stray default would @@ -159,8 +163,11 @@ export function apply(ctx: Context, config: Config): void { + 'VETOES the call; prefer plain notification events unless you intend to ' + 'intercept. (2) Never await something that only resolves after the current ' + 'turn (your code runs INSIDE a tool call of that turn — it would deadlock). ' - + '(3) The sandbox prevents accidental global pollution, not malice: `ctx` is ' - + 'the real, fully privileged runtime handle.', + + '(3) Your `ctx` is a restricted façade: you can register tools, observe ' + + 'events, provide/consume services, and use timers, but framework internals ' + + '(ctx.root, ctx.fiber, ctx.extend, ctx.plugin, …) are withheld. It is not a ' + + 'security boundary though — the services you inject (e.g. ctx.bash) reach the ' + + 'real runtime.', parameters: { code: { type: 'string', diff --git a/packages/cordis/tool-cordis/tests/mount.spec.ts b/packages/cordis/tool-cordis/tests/mount.spec.ts index 543b38307f..d0db8d7efa 100644 --- a/packages/cordis/tool-cordis/tests/mount.spec.ts +++ b/packages/cordis/tool-cordis/tests/mount.spec.ts @@ -386,13 +386,13 @@ describe('cordis_mount', () => { console.error('errored') const round = atob(btoa('hi')) const bytes = new TextEncoder().encode(round) - return { name: 'codec-' + new TextDecoder().decode(bytes), apply(ctx) { console.log('applied', typeof ctx.fiber) } } + return { name: 'codec-' + new TextDecoder().decode(bytes), apply(ctx) { console.log('applied', typeof ctx.on) } } `, }) expect(result.isError).toBe(false) expect(text(result)).toContain('plugin "codec-hi"') expect(log).toHaveBeenCalledWith('[cordis:dyn-1]', 'warned') - expect(log).toHaveBeenCalledWith('[cordis:dyn-1]', 'applied', 'object') + expect(log).toHaveBeenCalledWith('[cordis:dyn-1]', 'applied', 'function') expect(error).toHaveBeenCalledWith('[cordis:dyn-1]', 'errored') }) diff --git a/packages/cordis/tool-cordis/tests/sandbox-context.spec.ts b/packages/cordis/tool-cordis/tests/sandbox-context.spec.ts new file mode 100644 index 0000000000..50f5b441a2 --- /dev/null +++ b/packages/cordis/tool-cordis/tests/sandbox-context.spec.ts @@ -0,0 +1,156 @@ +import { describe, expect, it } from 'vitest' +import { call, setup, text } from './helpers.ts' + +/** + * The sandbox context façade is a whitelist, not a pass-through proxy: mount + * code reaches only the registration/eventing verbs, the timer helpers, a + * guarded `tools`, and its injected services. Every framework-plumbing member + * that could hand back an UNGUARDED context — through which a plugin could + * `ctx..tools.register({…})` to bypass the marker check and host-realm + * normalization — is denied. These are the regression guards for that escape + * class (the review finding on the original pass-through proxy). + */ + +/** Mount a plugin whose `apply` touches one framework member, and report the error text. */ +async function mountTouching(ctx: Awaited>, expr: string): Promise { + const result = await call(ctx, 'cordis_mount', { + code: `return { name: 'probe', inject: ['tools'], apply(ctx) { ${expr} } }`, + }) + expect(result.isError).toBe(true) + return text(result) +} + +describe('sandbox context façade — escape surface is closed', () => { + it.each([ + ['ctx.root', 'const c = ctx.root'], + ['ctx.parent', 'const c = ctx.parent'], + ['ctx.scope', 'const c = ctx.scope'], + ['ctx.fiber', 'const f = ctx.fiber'], + ['ctx.reflect', 'const r = ctx.reflect'], + ['ctx.registry', 'const r = ctx.registry'], + ['ctx.events', 'const e = ctx.events'], + ['ctx.extend()', 'ctx.extend({})'], + ['ctx.isolate()', 'ctx.isolate("x")'], + ['ctx.intercept()', 'ctx.intercept("x", {})'], + ['ctx.plugin()', 'ctx.plugin({ apply() {} })'], + ['ctx.set()', 'ctx.set("tools", 1)'], + ['ctx.mixin()', 'ctx.mixin("x", [])'], + ])('denies %s with a teaching error', async (_label, expr) => { + const ctx = await setup() + const message = await mountTouching(ctx, expr) + expect(message).toContain('sandbox ctx does not expose') + expect(message).toContain('withheld by design') + }) + + it('the classic ctx.root.tools.register bypass registers nothing and fails loud', async () => { + const ctx = await setup() + const result = await call(ctx, 'cordis_mount', { + code: ` + return { + name: 'root-bypass', + inject: ['tools'], + apply(ctx) { + ctx.root.tools.register({ + name: 'smuggled', + description: 'raw, unguarded', + parameters: { type: 'object', properties: {} }, + async execute() { return [] }, + }) + }, + } + `, + }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('sandbox ctx does not expose "root"') + // The whole point: the bypass never reaches the registry. + expect(ctx.tools.get('smuggled')).toBeUndefined() + }) + + it('rejects assignment to the façade rather than silently dropping it', async () => { + const ctx = await setup() + const result = await call(ctx, 'cordis_mount', { + code: 'return { name: \'writer\', apply(ctx) { ctx.stash = 1 } }', + }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('sandbox ctx is read-only') + }) + + it('denies a service whose method returns a Context (the .ctx escape), registering nothing', async () => { + // A cordis Service instance carries `.ctx` (a real Context), so + // `ctx.systemPrompt.ctx.root.tools.register(…)` would be a fresh unguarded + // handle. The service wrapper's return-value guard rejects any Context on + // the way back to sandbox code, so the escape never lands. (`systemPrompt` + // is in the setup harness, so the plugin activates and its apply runs.) + const ctx = await setup() + const result = await call(ctx, 'cordis_mount', { + code: ` + return { + name: 'svc-ctx-escape', + inject: ['systemPrompt', 'tools'], + apply(ctx) { + ctx.systemPrompt.ctx.root.tools.register({ + name: 'smuggled_via_service', + description: 'raw, unguarded', + parameters: { type: 'object', properties: {} }, + async execute() { return [] }, + }) + }, + } + `, + }) + expect(result.isError).toBe(true) + expect(text(result)).toContain('returned a cordis Context, which the sandbox does not expose') + expect(ctx.tools.get('smuggled_via_service')).toBeUndefined() + }) + + it('guards an async injected-service method: a host-realm Promise resolves through the guard', async () => { + // The return guard's Promise arm only fires for a HOST-realm Promise + // (a vm-realm one is not `instanceof` the host `Promise`). Provide a + // host-realm service from the test, then inject + await it from a mount: + // the resolved value is non-Context data and passes through. + const ctx = await setup() + ctx.plugin({ + name: 'host-async-svc', + apply(c) { c.provide('hostAsync', { grab: async () => 'host-fetched' }) }, + }) + await call(ctx, 'cordis_mount', { + code: ` + return { + name: 'async-consumer', + inject: ['hostAsync', 'tools'], + apply(ctx) { + harness.registerTool(ctx, harness.defineTool({ + name: 'do_fetch', + description: 'awaits the host async service', + parameters: {}, + async execute() { + const value = await ctx.hostAsync.grab() + return [{ type: 'text', text: value }] + }, + })) + }, + } + `, + }) + const result = await call(ctx, 'do_fetch', {}) + expect(result.isError).toBe(false) + expect(text(result)).toBe('host-fetched') + }) + + it('reads a symbol property as undefined and answers the `in` operator without throwing', async () => { + const ctx = await setup() + const result = await call(ctx, 'cordis_mount', { + code: ` + return { + name: 'introspector', + inject: ['tools'], + apply(ctx) { + const sym = ctx[Symbol.iterator] + console.log('probe', sym === undefined, 'tools' in ctx, 'on' in ctx, 'root' in ctx) + }, + } + `, + }) + expect(result.isError).toBe(false) + }) +})