Skip to content

fix: stabilize Claude bridge on OMP 17 #2

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
DevVig wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix: stabilize Claude bridge on OMP 17 #2

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
103 changes: 28 additions & 75 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,7 @@

<h1>omp-claude-bridge</h1>

<p><strong>Run Claude Code as a first-class model provider inside <a href="https://omp.sh">Oh My Pi</a> — with an AskClaude delegation tool and switchable 1M / 200K context windows.</strong></p>
<p><strong>Run Claude Code as a first-class model provider inside <a href="https://omp.sh">Oh My Pi</a> — with native OMP model selection and an AskClaude delegation tool.</strong></p>

<p>
<a href="LICENSE"><img alt="License: MIT" src="https://img.shields.io/badge/license-MIT-blue.svg"></a>
Expand All @@ -19,9 +19,9 @@

---

`omp-claude-bridge` lets you drive **Claude Code** — Opus, Sonnet, Haiku, and Fable — from inside Oh My Pi, with every tool call flowing through OMP's native TUI. It also exposes an **AskClaude** tool so any other provider can delegate a task or a second opinion to Claude Code, and it gives you **direct control over the context window** each model requests.
`omp-claude-bridge` lets you drive **Claude Code** — Opus, Sonnet, Haiku, and Fable — from inside Oh My Pi, with every tool call flowing through OMP's native TUI. It also exposes an **AskClaude** tool so any other provider can delegate a task or a second opinion to Claude Code.

Authentication and billing run through Claude Code and your Anthropic subscription via the official [Claude Agent SDK](https://github.com/anthropics/claude-agent-sdk-typescript) — this extension never stores credentials.
Authentication is delegated unchanged to the official [Claude Agent SDK](https://code.claude.com/docs/en/agent-sdk). The bridge provides no login flow and does not read, store, validate, log, or inject Claude credentials.

<div align="center">
<a href="assets/claude-bridge1.png"><img src="assets/claude-bridge1.png" width="49%"></a>&nbsp;
Expand All @@ -33,7 +33,8 @@ Authentication and billing run through Claude Code and your Anthropic subscripti
- [Features](#features)
- [Install](#install)
- [Quickstart](#quickstart)
- [Context window controls](#context-window-controls)
- [Authentication and billing](#authentication-and-billing)
- [Context handling](#context-handling)
- [Models](#models)
- [AskClaude tool](#askclaude-tool)
- [Configuration reference](#configuration-reference)
Expand All @@ -47,7 +48,7 @@ Authentication and billing run through Claude Code and your Anthropic subscripti

- **Claude Code as a provider** — pick Opus / Sonnet / Haiku / Fable from `/model`; tool calls render in OMP's TUI like any native provider.
- **AskClaude delegation tool** — from any other provider, hand a task or question to Claude Code (read-only, no-tools, or full read/write/bash), optionally in an isolated session.
- **Switchable context window** — force **1M** or **200K** globally, or leave it on measured per-model defaults. This is the headline addition in this fork.
- **Canonical OMP models** — one native picker entry per supported model, with OMP's catalog metadata preserved unchanged.
- **Session resume & persistence** — conversations survive across turns and reconnects.
- **Skills + AGENTS.md forwarding** — your OMP skills and context files are passed into Claude Code's system prompt.
- **Thinking support** — effort levels map through to Claude Code, including `xhigh` on Sonnet models.
Expand All @@ -73,7 +74,7 @@ omp plugin install ./omp-claude-bridge

</details>

Requires Oh My Pi (`omp`) and a working Claude Code login.
Requires Oh My Pi (`omp`) and either an existing local Claude Code login or `ANTHROPIC_API_KEY` in the parent environment.

## Quickstart

Expand All @@ -83,74 +84,33 @@ Requires Oh My Pi (`omp`) and a working Claude Code login.

To delegate from another provider instead, just ask: *"Ask Claude to review this plan and poke holes in it."*

## Context window controls
## Authentication and billing

Claude Code serves different context windows depending on the exact model id it receives (e.g. bare `claude-fable-5` serves 200K, while `claude-fable-5[1m]` serves 1M). `omp-claude-bridge` exposes both as **separate entries in the `/model` picker**, so you choose the window on demand:
The bridge calls the official Claude Agent SDK's `query()` without an `env` override, credential material, auth callback, or login command. Claude Code inherits the parent environment, remains the sole credential resolver, and applies its own authentication precedence. The bridge sets only non-credential runtime controls that suppress nonessential traffic, native Claude Code autocompaction, and Claude.ai cloud MCP auto-discovery.

- `claude-bridge/claude-opus-4-8` → **Opus 4.8 (1M)**
- `claude-bridge/claude-opus-4-8-200k` → **Opus 4.8 (200K)**
- An existing local Claude Code login currently draws from the user's Claude subscription quota.
- `ANTHROPIC_API_KEY` uses Anthropic API billing.

Switching window is just picking the other entry — no config edit, no reload. Every model appears once per window it supports, the `(1M)` / `(200K)` label is always shown, and each entry reports its true window so OMP's status bar and auto-compaction threshold stay accurate.
The bridge has no login flow and does not handle credentials. Anthropic's [June 15 support update](https://support.claude.com/en/articles/15036540-use-the-claude-agent-sdk-with-your-claude-plan) currently says third-party Agent SDK usage draws from subscription limits. The [Agent SDK overview](https://code.claude.com/docs/en/agent-sdk) also says third-party developers may not offer Claude.ai login or rate limits without prior approval. This repository does not claim explicit Anthropic approval; policy and billing behavior are subject to Anthropic change.

### Default window

The **unsuffixed** id (e.g. `claude-opus-4-8`) maps to a default window; the other window gets a `-1m` / `-200k` suffixed id. `provider.contextWindow` in `~/.omp/agent/claude-bridge.json` picks that default — it no longer hides models, it only decides which window is unsuffixed:
## Context handling

```json
{
"provider": {
"contextWindow": "auto"
}
}
```

| Mode | Default (unsuffixed) window |
| ---- | -------- |
| `"auto"` *(default)* | Per-model measured default. Respects `plan` and `longContextExtraUsage`. |
| `"1m"` | 1M where the model has a 1M runtime, else its only window. |
| `"200k"` | 200K where the model has a 200K runtime, else its only window. |
The bridge registers exactly one canonical OMP catalog entry per supported model and passes that model ID unchanged to Claude Code. It does not create context-window aliases, rewrite model IDs, or advertise account-entitlement variants.

Both windows stay in the picker regardless of this setting (wherever a runtime exists); it only changes which one is the plain, unsuffixed id. So `modelRoles` / `enabledModels` that reference `claude-bridge/claude-opus-4-8` keep working and follow the default.

### Windows offered per model

| Model | 200K entry | 1M entry | `auto` default |
| ----- | :--------: | :------: | :------------: |
| `claude-opus-4-8` | ✓ | ✓ | 1M |
| `claude-opus-4-7` | — | ✓ | 1M |
| `claude-opus-4-6` | ✓ | ✓ | 200K¹ |
| `claude-fable-5` | ✓ | ✓ | 200K |
| `claude-sonnet-5` | ✓ | ✓ | 1M |
| `claude-sonnet-4-6` | ✓ | ✓ | 200K² |
| `claude-haiku-4-5` | ✓ | — | 200K |

¹ Opus 4.6's `auto` default is 1M when `plan: "max"` or `longContextExtraUsage: true`.
² Sonnet 4.6's `auto` default is 1M when `longContextExtraUsage: true`.

The suffixed alternate exists only for the window that isn't the default — e.g. under `auto` you get `claude-opus-4-8` (1M) + `claude-opus-4-8-200k`, and under `"200k"` you get `claude-opus-4-8` (200K) + `claude-opus-4-8-1m`.

> Forcing 1M is a *request*: some models may still be **served** 200K by your subscription entitlement. Set `CLAUDE_BRIDGE_DEBUG=1` to log the served window (see [Debugging](#debugging)).

> An invalid `contextWindow` value logs a warning and falls back to `"auto"`, so a typo never breaks startup.
Use [`omp-ccwindow`](https://github.com/DevVig/omp-ccwindow) when you want to lower OMP's active context value for status display and compaction timing. That plugin changes OMP's local model clone; it does not change the Claude API model's native context capacity.

## Models

Pick any of these from `/model` — each entry shows a `(1M)` or `(200K)` label. The exact ids below assume the default `contextWindow: "auto"`; which id is unsuffixed vs `-1m` / `-200k` follows your configured [default window](#default-window).

| Picker id (auto) | Window |
| --------- | ------ |
| `claude-bridge/claude-fable-5` | 200K |
| `claude-bridge/claude-fable-5-1m` | 1M |
| `claude-bridge/claude-opus-4-8` | 1M |
| `claude-bridge/claude-opus-4-8-200k` | 200K |
| `claude-bridge/claude-opus-4-7` | 1M |
| `claude-bridge/claude-opus-4-6` | 200K (1M on Max / Extra Usage) |
| `claude-bridge/claude-opus-4-6-1m` | 1M |
| `claude-bridge/claude-sonnet-5` | 1M (supports `xhigh`) |
| `claude-bridge/claude-sonnet-5-200k` | 200K |
| `claude-bridge/claude-sonnet-4-6` | 200K (supports `xhigh`) |
| `claude-bridge/claude-sonnet-4-6-1m` | 1M |
| `claude-bridge/claude-haiku-4-5` | 200K (cheapest) |
The bridge exposes these canonical picker entries:

- `claude-bridge/claude-fable-5`
- `claude-bridge/claude-opus-4-8`
- `claude-bridge/claude-opus-4-7`
- `claude-bridge/claude-opus-4-6`
- `claude-bridge/claude-sonnet-5`
- `claude-bridge/claude-sonnet-4-6`
- `claude-bridge/claude-haiku-4-5`

Bash commands issued by Claude Code get a 120-second default timeout (matching Claude Code's default), since OMP's bash has no timeout by default.

Expand Down Expand Up @@ -188,9 +148,6 @@ Config is read from `~/.omp/agent/claude-bridge.json` (global) and the project O
"defaultIsolated": false
},
"provider": {
"contextWindow": "auto",
"plan": "pro",
"longContextExtraUsage": false,
"strictMcpConfig": true
}
}
Expand All @@ -213,17 +170,14 @@ Config is read from `~/.omp/agent/claude-bridge.json` (global) and the project O

| Key | Default | Description |
| --- | ------- | ----------- |
| `contextWindow` | `"auto"` | `"auto"`, `"1m"`, or `"200k"`. See [Context window controls](#context-window-controls). |
| `plan` | `"pro"` | Set to `"max"` to enable Opus 4.6 at 1M in `auto`. |
| `longContextExtraUsage` | `false` | Opt into metered 1M usage (enables Sonnet 4.6 1M everywhere, Opus 4.6 1M on Pro). |
| `appendSystemPrompt` | `true` | Append OMP's AGENTS.md and skills. |
| `settingSources` | — | Claude Code filesystem settings to load; only applied when `appendSystemPrompt: false`. |
| `strictMcpConfig` | `true` | Block MCP servers from `~/.claude.json` / `.mcp.json`. Cloud MCP is always blocked. |
| `strictMcpConfig` | `true` | Ignore filesystem MCP configuration outside the bridge's OMP tool server; Claude.ai cloud MCP auto-discovery is also disabled. |
| `pathToClaudeCodeExecutable` | — | Path to the `claude` binary, if the bundled one can't run on your OS/filesystem. |

## How it works

OMP's built-in tools are bridged to Claude Code and back, so from your side it behaves like any other OMP provider. Model routing lives in [`src/models.ts`](src/models.ts), which is deliberately free of runtime imports so the context-window policy stays unit-testable in isolation. On registration, the extension projects the pi-ai model list, applies the selected context-window policy, and registers the resulting models with OMP.
OMP's built-in tools are bridged to Claude Code and back, so from your side it behaves like any other OMP provider. On every extension initialization, the bridge registers the canonical OMP Anthropic catalog entries while retaining the parent session's process-global stream function. Agent SDK queries receive the selected canonical model ID and resolve authentication entirely inside Claude Code.

## Debugging

Expand Down Expand Up @@ -251,12 +205,11 @@ See [CONTRIBUTING.md](CONTRIBUTING.md) for the full workflow. CI runs typecheck

- Original **[pi-claude-bridge](https://github.com/elidickinson/pi-claude-bridge)** by **[Eli Dickinson](https://github.com/elidickinson)** — the streaming provider, MCP/tool bridging, session resume, and AskClaude tool this project builds on.
- Initial inspiration from [claude-agent-sdk-pi](https://github.com/prateekmedia/claude-agent-sdk-pi) by Prateek Sunal.
- **Oh My Pi port and context-window controls** by **[Jonathan Borgwing](https://github.com/DevVig)**.
- **Oh My Pi port and maintenance** by **[Jonathan Borgwing](https://github.com/DevVig)**.

See [NOTICE](NOTICE) for full attribution.

> Anthropic [announced and then unannounced](https://support.claude.com/en/articles/15036540-use-the-claude-agent-sdk-with-your-claude-plan) a change to how Agent-SDK tool usage is billed. As of June 15, 2026 it uses your subscription quota just like Claude Code direct.

## License

[MIT](LICENSE) © 2026 Eli Dickinson (original) and Jonathan Borgwing (Oh My Pi port and context-window controls).
[MIT](LICENSE) © 2026 Eli Dickinson (original) and Jonathan Borgwing (Oh My Pi port).
Loading
Loading