customizations-in-the-agent-host

$npx mdskill add microsoft/vscode/customizations-in-the-agent-host

Customizations are the plugins, MCP servers, agents, skills and instructions that an agent-host session exposes to a model. This skill covers **enablement**: how a customization's on/off state is decided, stored, published and enforced.

SKILL.md

.github/skills/customizations-in-the-agent-hostView on GitHub ↗
---
name: customizations-in-the-agent-host
description: Architecture and hard-won debugging lessons for customization enablement (plugins, MCP servers, agents, skills, instructions) in the agent host. Use when changing how customizations are discovered, published, enabled/disabled, or handed to a provider SDK; when a customization shows the wrong enabled state in the UI; or when a disabled MCP server or plugin is still reaching the model.
---

# Customizations in the Agent Host

Customizations are the plugins, MCP servers, agents, skills and instructions that an agent-host session exposes to a model. This skill covers **enablement**: how a customization's on/off state is decided, stored, published and enforced.

Read this before changing anything in:

- `src/vs/platform/agentHost/node/agentHostCustomizationEnablementService.ts`
- `src/vs/platform/agentHost/node/shared/customizationEnablementGate.ts`
- `src/vs/platform/agentHost/common/customizationEnablement.ts`
- the per-provider customization paths under `node/copilot/`, `node/claude/`, `node/codex/`

## The mental model

### Enablement is a list of scoped decisions, not a boolean

A plugin or MCP server carries `enablement?: CustomizationEnablement[]` — an ordered list of decisions, most specific first:

| Kind | Owner | Meaning |
|---|---|---|
| `Session` | agent host | this session only |
| `Workspace` | agent host | this working directory (carries `uri`) |
| `Global` | client, unless the host discovered it | everywhere |

`isCustomizationEnabled()` picks the first entry; absent entries inherit from the next scope out, and an absent list means enabled. Never re-add a parallel `enabled` boolean to plugins or MCP servers — a second source of truth is what this design exists to remove. (`DirectoryCustomization` legitimately keeps a plain `enabled`; consumers derive its state through `isCustomizationEnabled()` all the same.)

### Who owns `Global`

The host owns `Global` for anything it **discovered**, even when that thing lives inside a plugin the client forwarded. The client owns `Global` only for what it actually **bundled**. The discriminator is `childEnablement` — *not* `clientId`. A plugin whose `childEnablement` record has a key for a child MCP server means "the client bundled this child and owns its global decision"; anything else the host discovered by parsing the plugin directory.

### The overlay model

This is the single most important invariant, and getting it wrong caused a self-defeating write that silently erased user decisions:

> The client's value is the **base**. Host decisions layer **on top**. "Store only what differs" compares against that base — never against a hardcoded `true`.

Concretely, in `agentHostCustomizationEnablementService.ts`:

- `_clientGlobalEnablement` is the in-memory **base**, per session. `_setClientGlobal` writes here and **must never touch `_persistent`**.
- `_persistent` holds only host-owned decisions that differ from the base.
- Inheritance runs host → base → `DEFAULT_CUSTOMIZATION_ENABLED` (`_globalEnablement`, `_workspaceInheritedEnablement`).
- `_setPersistentDecision` deletes a decision that matches its inherited value, so the UI can restore inheritance without a third "Inherit" action.

Keeping these separate is what stops this loop: `replaceEnablement` → synchronous republish → client's stale `[{global,true}]` → `_setClientGlobal` → "store only what differs" deletes the entry just written.

### Identity must not depend on tree position

An MCP server's storage key is derived from `owningPluginUri`, never from where it happened to appear in the published tree. The same server discovered top-level (`mcpServers#azure`) and nested under a plugin (`<pluginURI>#mcp=azure`) must produce **one** key, or a decision written in one session is invisible to the next. `targetForMcpServer()` and `withOwningPluginUri()` in the gate are the only supported way to build that target.

### Pending means fail closed — but pending must be transient

`resolve()` returns `{kind:'pending'}` when the session isn't initialized yet (`reason: 'session'`) or the working directory isn't known (`reason: 'workingDirectory'`). `resolveCustomizationEnablement()` records those in `pendingCustomizationIds`, and `isCustomizationSdkEligible()` makes SDK-boundary consumers **fail closed**.

Two rules follow:

1. Never "fix" a pending customization by defaulting it to enabled.
2. Never cache the pending verdict. Pending is a *temporary* state; recompute it, and republish when the enablement service settles. A cached `pendingEnablement: true` silently drops a plugin forever.

A pending child MCP server also marks its **containing plugin** pending, because a plugin directory can cause the SDK to discover the child on its own.

## The rule that prevents most bugs

> **Every publication, emit and SDK-handoff path must go through resolution.**

Three separate production bugs on this feature came from paths that skipped it (four sites in `copilotAgent`, one in `mcpCustomizationController`, plus the post-startup reconcile). When you add a new place that publishes a customization or hands one to a provider, route it through `resolveCustomizationEnablement()` — do not read a stored snapshot.

The corollary: **state computed once must be rebuilt when its inputs change.** Merged per-client enablement maps that are only appended to will leak stale decisions onto a newly parsed plugin with the same URI.

## Enforcement is per-provider, and they differ

| Provider | Launch | Mid-session |
|---|---|---|
| Copilot | `disabledMcpServers` on `SessionConfig` at open/resume | `rpc.mcp.enable` / `rpc.mcp.disable` |
| Claude | session options at materialize | live reconcile |
| Codex | MCP servers supplied only at `thread/start` / resume | **not possible** — documented limitation |

Enforcement traps:

- The post-startup reconcile runs ~800 ms after launch. Anything gated *only* there can still start, and request authentication, during the launch window. Gate at launch too.
- The SDK's `not_configured` status must **not** be translated as `enabled: true`. That bug made a correct "skip this disabled server" decision look like "enable it".
- Copilot launch gating must include names from user `mcp.config.list`, not only the client snapshot. Reconcile and authentication also resolve newly discovered names before the host retains them; later publication must not turn an inherited enabled default into explicit intent to enable an SDK-configured-disabled server.
- Root MCP customization IDs use the owning AHP session ID, not its SDK backing ID. On restoration, derive declared MCP names from the applied root/plugin definitions, never from the retained host-published runtime inventory.
- Connection lifecycle (`pending`, `connected`, `stopped`) is not enablement intent. Confirm external toggles against passive `mcp.listConfigured` enablement, ignoring stale asynchronous snapshots and host-requested changes.

Client-provided MCP servers (ones the client bundled rather than the host discovering) must route their global Disable to the **client**, not the host — the host correctly refuses to store a client-owned global, so routing it to the host is a silent no-op. `isClientBundled` is the discriminator; see `getBuiltinMcpServerEnablementActions` in `mcpListWidget.ts`.

## Debugging playbook

Rank your signals. They disagree, and the model lies:

1. **Primary** — the published customization object and the host's stored decision.
2. **Definitive for tool exposure** — `tool_search override … clientMatched=[…]` in `Agent Host.log`.
3. **Secondary** — the customizations UI.
4. **Corroborating only** — what the model says. A model answered "yes, I have azure tools" while the server was `stopped`. Never conclude from this alone.

Reproduction recipe that exercises the real failure surface: start a session, send a message to boot everything, disable the thing, then ask a **new** session whether it has those tools. Cross-session persistence is where identity and overlay bugs surface.

Use the `agent-host-logs` skill for exported bundles, and the `launch` skill to drive a real instance.

Known false alarms:

- The `azure` plugin contributes 24 `azure-*` **skills** in addition to its MCP server. Disabling the server correctly leaves the skills enabled.
- `sleepy` is a deliberately broken test server that cannot initialize — useless as a positive control.

## Gotchas that cost real time

- `CustomizationType` and `CustomizationEnablementKind` are `const enum`s. A **type-only import compiles and lints cleanly, then crashes at runtime.** Import them as values.
- The unit-test runner uses pre-compiled output and does **not** fail on type errors. Always run `npm run typecheck-client` separately.
- Integration-test servers start with `--quiet`. Wiring that you hoist out of `if (!options.quiet)` in `agentHostServerMain.ts` silently changes their behaviour — hoisting `setWorktreeIsolation` made quiet servers resolve `isolation: 'folder'` and broke seven unrelated tests.
- Protocol files under `common/state/protocol/` carry a "DO NOT EDIT — auto-generated" banner. Changes there need a matching backport to the `agent-host-protocol` repo.
- Electron runner setup aborts (`EEXIST`/`EPERM`/`ENOENT`/ReactiveObjC) are environmental. Recover with `rm -rf .build/electron && npm run electron`.

## Validation

```sh
npm run typecheck-client
./scripts/test.sh --grep "customizationEnablement"
./scripts/test-integration.sh --runGlob "**/agentHost/**/{sessionConfig,toolApproval,codexCustomizations}.integrationTest.js"
./scripts/test-integration.sh --run src/vs/platform/agentHost/test/node/claudeAgent.integrationTest.ts
```

Tests that construct an agent session directly need `IAgentHostCustomizationEnablementService` in their service collection; use `createNoopCustomizationEnablementService()` from `src/vs/platform/agentHost/test/node/testCustomizationEnablementService.ts`.

More from microsoft/vscode

SkillDescription
accessibilityPrimary accessibility skill for VS Code. REQUIRED for new feature and contribution work, and also applies to updates of existing UI. Covers accessibility help dialogs, accessible views, verbosity settings, signals, ARIA announcements, keyboard navigation, and ARIA labels/roles.
act-on-feedbackAct on user feedback attached to the current session. Use when the user submits feedback on the session's changes via the Submit Feedback button.
agent-customization**WORKFLOW SKILL** — Create, update, review, fix, or debug VS Code agent customization files (.instructions.md, .prompt.md, .agent.md, SKILL.md, copilot-instructions.md, AGENTS.md). USE FOR: saving coding preferences; troubleshooting why instructions/skills/agents are ignored or not invoked; configuring applyTo patterns; defining tool restrictions; creating custom agent modes or specialized workflows; packaging domain knowledge; fixing YAML frontmatter syntax. DO NOT USE FOR: general coding questions (use default agent); runtime debugging or error diagnosis; MCP server configuration (use MCP docs directly); VS Code extension development. INVOKES: file system tools (read/write customization files), ask-questions tool (interview user for requirements), subagents for codebase exploration. FOR SINGLE OPERATIONS: For quick YAML frontmatter fixes or creating a single file from a known pattern, edit the file directly — no skill needed.
agent-host-chat-contributionsBuild and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.
agent-host-e2e-testsUse when writing, recording, updating, validating, or troubleshooting Agent Host end-to-end and real-provider integration tests under src/vs/platform/agentHost/test/node/e2e and providerIntegration. Covers adding a cross-provider test, re-recording fixtures after an SDK bump, required cross-platform Azure validation, gating non-deterministic or platform-specific tests, and diagnosing replay cache misses.
agent-host-logsAnalyze Agent Host debug log exports. Use when given an ah-logs or ahp-logs zip/folder, an Export Agent Host Debug Logs bundle, events.jsonl, AHP JSONL transport logs, Agent Host.log, remote-agenthost.log, or copilot-logs.
author-contributionsIdentify all files a specific author contributed to on a branch vs its upstream, tracing code through renames. Use when asked who edited what, what code an author contributed, or to audit authorship before a merge. This skill should be run as a subagent — it performs many git operations and returns a concise table.
auto-perf-optimizeRun agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.
azure-pipelinesUse when validating Azure DevOps pipeline changes for the VS Code build. Covers queueing builds, checking build status, viewing logs, and iterating on pipeline YAML changes without waiting for full CI runs.
build-champ-triageFind out why a build has failed