dsh-llm-github-copilot

dsh-llm-github-copilot

English | 中文

GitHub Copilot sign-in for DeepSeek Harness.

Sign in with your GitHub account and use every Copilot model your plan includes — GPT-5 family, Claude, Gemini, and more — inside DeepSeek Harness.

The harness already ships a github-copilot provider that can serve those models; what it cannot do is sign you in. This plugin fills exactly that gap: it runs the GitHub device flow and publishes the credential that provider authenticates from. Requests, model discovery, images, and streaming are all handled by the harness route, not here. See docs/adr/0002-narrow-to-credential-provider.md.

Requirements

• DeepSeek Harness 0.1.7-rc.1 through 0.1.7-rc.2, or 0.2.0-rc.1 from npm's next channel, with its bundled @deepseek-ai/dsh-llm-pi-ai route (mounted by default). This plugin writes the credential that route reads. This baseline uses Loader-owned volatile configuration and forwards credential reference commits to the Web client, so the sign-in dialog closes immediately once the device flow completes.
• A GitHub account with a Copilot subscription.
• Node.js ≥ 24.

Install

From npm (recommended)

dsh plugin --profile web add @lujianjun19/dsh-llm-github-copilot

From GitHub

dsh plugin --profile web add github:lujianjun19/dsh-llm-github-copilot

pnpm 10 and newer block git dependency build scripts by default. If installation asks you to approve the package build, add the exact package key it reports to the profile's pnpm-workspace.yaml:

allowBuilds:
  dsh-llm-github-copilot: true

Then repeat the install command.

After installation, register the plugin in the profile's cordis.patch.yml:

- insert:
    - id: llm-github-copilot
      name: '@lujianjun19/dsh-llm-github-copilot'

Restart DSH to activate:

dsh web

Update

The DSH profile uses pnpm internally, and its lockfile pins the exact installed version. Re-running dsh plugin add alone may not pick up a newer release. Use either of the following commands from any directory:

# Using npm (simplest — no extra flags needed)
npm install --prefix ~/.dsh/profiles/web @lujianjun19/dsh-llm-github-copilot@latest

# Using pnpm
pnpm add --dir ~/.dsh/profiles/web @lujianjun19/dsh-llm-github-copilot@latest --no-frozen-lockfile

Then restart DSH:

dsh web

To confirm the installed version, open Settings → GitHub Copilot in the Harness browser UI — the version number is shown at the bottom of the panel.

Sign in

Run the login command inside the Harness chat:

/copilot-login

Follow the instructions — open the verification URL in your browser, enter the displayed code, and authorize the GitHub App. The token is stored automatically once you complete authorization. Run /copilot-status to confirm the connection and see the available models.

To sign out:

/copilot-logout

You can also set the token directly as an environment variable (useful for CI or headless setups):

export GITHUB_COPILOT_OAUTH_TOKEN=<your-github-oauth-token>

Supported token types:

Prefix	Source
gho_	OAuth token (gh auth login)
github_pat_	Fine-grained PAT (requires Copilot permission)
ghu_	GitHub App user token (VS Code client)

ghp_ classic PATs are not accepted by the Copilot API.

Features

Device-flow sign-in — no token to create or paste. Run /copilot-login, or use the Settings page, and authorize in the browser; the credential is stored automatically.

Automatic refresh, handled downstream — the credential carries your long-lived GitHub token. The harness route exchanges it for a short-lived Copilot token, refreshes it, and derives your account's API endpoint (individual, business, or enterprise) on its own.

Ambient token adoption — an existing GITHUB_COPILOT_OAUTH_TOKEN, whether exported or stored from an earlier version of this plugin, is adopted on start so you are not asked to sign in twice.

Settings page — a dedicated GitHub Copilot section in the Harness Web settings UI (gear icon → GitHub Copilot) for signing in, checking status, and signing out. Models are selected under Settings → Models, in the github-copilot provider.

Slash commands — /copilot-login, /copilot-status, and /copilot-logout for surfaces without the settings UI.

Selecting a model

Signing in publishes the credential; it does not add a model provider by itself. The harness's pi-ai route ships github-copilot in its installed catalog, but that route mounts dormant — with an empty provider list — until you explicitly add it once:

1. Open Settings → Models (a different page from this plugin's own Settings → GitHub Copilot sign-in page).
2. Choose Add provider, then pick github-copilot from the catalog list. It is a built-in pi-ai catalog entry, so no endpoint or API key field is needed — the plugin's stored credential covers that.
3. Save. The harness performs the Copilot token exchange and discovers your account's available models the first time the route is used or its catalog is requested; this can take a few seconds.
4. Reopen the model selector in the chat input box — the GitHub Copilot models now appear there.

Until step 2 is done, the chat input's model selector shows no GitHub Copilot models at all, regardless of sign-in state; /copilot-status still reports authenticated because that reflects the stored credential, not whether a model route was added.

Upgrading from 0.4.x

Versions up to 0.4.5 registered their own github-copilot-official provider. That provider no longer exists. If your settings or sessions name it, switch them to the github-copilot provider — the model ids are the same. Your stored credential is adopted automatically; you do not need to sign in again.

Configure

The plugin works with no configuration. The only setting is which credential reference holds the GitHub OAuth token:

- id: llm-github-copilot
  config:
    oauthTokenEnv: GITHUB_COPILOT_OAUTH_TOKEN   # credential reference / env var

Everything about models, endpoints, and request behaviour is configured on the harness route under the llm-pi-ai profile entry, not here. On Harness 0.1.7-rc.1 and newer, live settings are stored in the active profile's cordis.patch.yml, rather than in a separate settings.yaml section.

Develop

Requires Node.js ≥ 24 and npm.

cd /path/to/dsh-llm-github-copilot
npm run build      # concatenate source fragments → lib/index.js and lib/client.js
npm test           # build + deterministic artifact, i18n, and metadata tests
npm run check      # build + tests + npm pack dry-run
npm run deploy     # test → build → atomic deploy with rollback backup

Install the checkout into a local profile directly:

dsh plugin --profile web add .

After Host changes restart DSH; after Client-only changes a hard refresh (Ctrl+Shift+R) is usually enough.

Rollback

npm run deploy keeps a timestamped backup before each install. To roll back, stop DSH and restore the wanted backup:

cp -r ~/.dsh/plugin-backups/dsh-llm-github-copilot/<timestamp> \
      ~/.dsh/profiles/web/node_modules/@lujianjun19/dsh-llm-github-copilot
dsh web

Troubleshooting

Model selector shows no GitHub Copilot group after restart The plugin loaded in a previous process that didn't pick up the new package. Restart dsh web. The stored token is preserved; no need to sign in again.

Models appear but Claude is missing Your egress IP is restricted. Export HTTPS_PROXY pointing to a proxy that exits from an unrestricted region, then restart dsh web.

/copilot-login times out or reports a network error Transient network issue during device-code polling. Run /copilot-login again to get a fresh code (the old one is invalidated automatically).

Signed in (/copilot-status shows authenticated), but the chat model selector shows no GitHub Copilot models at all Signing in only publishes the credential; it does not add a model route by itself. Open Settings → Models → Add provider and add github-copilot from the catalog list (see Selecting a model). Until that provider is added, the harness's llm-pi-ai route stays dormant with an empty provider list and offers no models, even though sign-in succeeded.

configurable provider "github-copilot" is already declared This plugin registers no provider of its own — it publishes a credential for the harness's github-copilot provider. If no models appear, confirm you signed in (/copilot-status) and that the github-copilot provider is added under Settings → Models. Verify that your cordis.patch.yml uses id: llm-github-copilot and name: '@lujianjun19/dsh-llm-github-copilot'.

Token expired No action needed. The plugin stores the long-lived GitHub OAuth token and refreshes the short-lived Copilot API token automatically before it expires. Only an explicit sign-out or token revocation requires a new /copilot-login.

License

MIT