feat: open-swe dashboard for per-user profile config (#1302)
* feat: dashboard backend — GitHub OAuth, profile CRUD, admin endpoints
Adds agent/dashboard/ FastAPI router mounted at /dashboard/api covering:
- GitHub App OAuth login → JWT cookie session (cross-domain ready)
- profile CRUD against LangGraph Store with model+effort validation
- admin gate via CONFIGURED_ADMINS
- /repos via /user/installations using the user's encrypted OAuth token
CORS allowlist on webapp.py is opt-in via DASHBOARD_ALLOWED_ORIGINS so the
Vercel-hosted frontend can call the LangSmith deployment with credentials.
* feat: apply dashboard profile model/effort overrides in get_agent
Look up the triggering user's GitHub login from config (direct field or
GITHUB_USER_EMAIL_MAP reverse lookup), read their profile from the Store,
and apply default_model + reasoning_effort to make_model when both are
valid. Effort 'max' is captured on the profile but not yet wired through —
the OpenAI Reasoning Literal doesn't accept it.
* feat: ui/ TanStack Start dashboard for profile config
Scaffolded with the shadcn b7CScJIjA preset (TanStack Start template,
base-ui primitives, Tailwind v4). Three routes:
- /login — Sign in with GitHub (links to /dashboard/api/auth/login)
- /profile — Edit default model, reasoning effort, default repo
- /admin — Admin-only: list users and edit other profiles
API client (src/lib/api.ts) uses credentials: include so the osw_session
cookie set by the OAuth callback rides cross-origin. VITE_DASHBOARD_API_BASE_URL
points at the LangSmith deployment.
Effort options re-render when the model changes; 'max' on Opus 4.7 is
captured on the profile but ignored downstream until anthropic reasoning
is wired through make_model.
* feat: searchable Combobox for default repo picker
Replaces the Select with a base-ui Combobox so users can filter by typing,
the popup is wider than the trigger so full owner/repo names are readable,
and the list caps at max-h-80 to stay on screen.
* fix: address review comments + wire default_repo and Anthropic thinking
Security/correctness fixes from PR review:
* Open redirect: validate `redirect_to` in `/auth/login` against
`DASHBOARD_BASE_URL` + `DASHBOARD_ALLOWED_ORIGINS` before signing it
into the state JWT. Anything off-allowlist falls back to the dashboard
base URL. (PR #1302 r3250054386)
* Login CSRF: bind the OAuth `state` to the requesting browser. At
`/auth/login` we generate a fresh nonce, set it as a short-lived
HttpOnly SameSite=Lax cookie scoped to `/dashboard/api/auth`, and
embed `hash_state_nonce(nonce)` in the state JWT. At `/auth/callback`
we require the cookie nonce to hash-match the state JWT's nonce_hash
(constant-time compare). (PR #1302 r3250054395)
* RMW race in profile vs token writes: split storage into two
namespaces — `["profiles"]` for user-editable settings and
`["oauth_tokens"]` for the encrypted GitHub token. Each upsert now
only writes its own namespace so an in-flight profile save can no
longer clobber a fresh token from a concurrent re-login (and vice
versa). (PR #1302 r3250054393)
* /repos pagination: follow `Link: rel="next"` for both
`/user/installations` and per-installation `/repositories` with
per_page=100, capped at 1000 items. (PR #1302 r3250054401)
Feature wires:
* default_repo: applied as a fallback in `get_slack_repo_config` (after
explicit-repo / thread metadata, before the env defaults) and in the
Linear webhook (after comment-body extraction, before team mapping).
Both paths resolve the triggering user's GitHub login via
GITHUB_USER_EMAIL_MAP and read the profile's default_repo.
* Anthropic "thinking" effort: `make_model` now accepts a `thinking`
kwarg; `get_agent` maps profile effort {low,medium,high,xhigh,max}
to budget_tokens {1k,4k,12k,32k,60k} when the chosen model is
anthropic. OpenAI path still ignores "max" since the Literal doesn't
accept it.
2026-05-15 11:23:53 -07:00
|
|
|
"""GitHub App OAuth code-exchange and signed-JWT session cookie."""
|
|
|
|
|
|
|
|
|
|
from __future__ import annotations
|
|
|
|
|
|
|
|
|
|
import hashlib
|
|
|
|
|
import hmac
|
|
|
|
|
import logging
|
|
|
|
|
import os
|
|
|
|
|
import secrets
|
|
|
|
|
import time
|
2026-05-21 10:44:04 -07:00
|
|
|
from datetime import UTC, datetime, timedelta
|
feat: open-swe dashboard for per-user profile config (#1302)
* feat: dashboard backend — GitHub OAuth, profile CRUD, admin endpoints
Adds agent/dashboard/ FastAPI router mounted at /dashboard/api covering:
- GitHub App OAuth login → JWT cookie session (cross-domain ready)
- profile CRUD against LangGraph Store with model+effort validation
- admin gate via CONFIGURED_ADMINS
- /repos via /user/installations using the user's encrypted OAuth token
CORS allowlist on webapp.py is opt-in via DASHBOARD_ALLOWED_ORIGINS so the
Vercel-hosted frontend can call the LangSmith deployment with credentials.
* feat: apply dashboard profile model/effort overrides in get_agent
Look up the triggering user's GitHub login from config (direct field or
GITHUB_USER_EMAIL_MAP reverse lookup), read their profile from the Store,
and apply default_model + reasoning_effort to make_model when both are
valid. Effort 'max' is captured on the profile but not yet wired through —
the OpenAI Reasoning Literal doesn't accept it.
* feat: ui/ TanStack Start dashboard for profile config
Scaffolded with the shadcn b7CScJIjA preset (TanStack Start template,
base-ui primitives, Tailwind v4). Three routes:
- /login — Sign in with GitHub (links to /dashboard/api/auth/login)
- /profile — Edit default model, reasoning effort, default repo
- /admin — Admin-only: list users and edit other profiles
API client (src/lib/api.ts) uses credentials: include so the osw_session
cookie set by the OAuth callback rides cross-origin. VITE_DASHBOARD_API_BASE_URL
points at the LangSmith deployment.
Effort options re-render when the model changes; 'max' on Opus 4.7 is
captured on the profile but ignored downstream until anthropic reasoning
is wired through make_model.
* feat: searchable Combobox for default repo picker
Replaces the Select with a base-ui Combobox so users can filter by typing,
the popup is wider than the trigger so full owner/repo names are readable,
and the list caps at max-h-80 to stay on screen.
* fix: address review comments + wire default_repo and Anthropic thinking
Security/correctness fixes from PR review:
* Open redirect: validate `redirect_to` in `/auth/login` against
`DASHBOARD_BASE_URL` + `DASHBOARD_ALLOWED_ORIGINS` before signing it
into the state JWT. Anything off-allowlist falls back to the dashboard
base URL. (PR #1302 r3250054386)
* Login CSRF: bind the OAuth `state` to the requesting browser. At
`/auth/login` we generate a fresh nonce, set it as a short-lived
HttpOnly SameSite=Lax cookie scoped to `/dashboard/api/auth`, and
embed `hash_state_nonce(nonce)` in the state JWT. At `/auth/callback`
we require the cookie nonce to hash-match the state JWT's nonce_hash
(constant-time compare). (PR #1302 r3250054395)
* RMW race in profile vs token writes: split storage into two
namespaces — `["profiles"]` for user-editable settings and
`["oauth_tokens"]` for the encrypted GitHub token. Each upsert now
only writes its own namespace so an in-flight profile save can no
longer clobber a fresh token from a concurrent re-login (and vice
versa). (PR #1302 r3250054393)
* /repos pagination: follow `Link: rel="next"` for both
`/user/installations` and per-installation `/repositories` with
per_page=100, capped at 1000 items. (PR #1302 r3250054401)
Feature wires:
* default_repo: applied as a fallback in `get_slack_repo_config` (after
explicit-repo / thread metadata, before the env defaults) and in the
Linear webhook (after comment-body extraction, before team mapping).
Both paths resolve the triggering user's GitHub login via
GITHUB_USER_EMAIL_MAP and read the profile's default_repo.
* Anthropic "thinking" effort: `make_model` now accepts a `thinking`
kwarg; `get_agent` maps profile effort {low,medium,high,xhigh,max}
to budget_tokens {1k,4k,12k,32k,60k} when the chosen model is
anthropic. OpenAI path still ignores "max" since the Literal doesn't
accept it.
2026-05-15 11:23:53 -07:00
|
|
|
from typing import Any
|
|
|
|
|
from urllib.parse import urlparse
|
|
|
|
|
|
|
|
|
|
import httpx
|
|
|
|
|
import jwt
|
|
|
|
|
from fastapi import HTTPException, Request
|
|
|
|
|
|
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
|
|
|
|
COOKIE_NAME = "osw_session"
|
|
|
|
|
STATE_COOKIE_NAME = "osw_oauth_state"
|
|
|
|
|
SESSION_TTL_SECONDS = 7 * 24 * 60 * 60
|
|
|
|
|
STATE_TTL_SECONDS = 600
|
|
|
|
|
JWT_ALG = "HS256"
|
|
|
|
|
|
|
|
|
|
GITHUB_APP_CLIENT_ID = os.environ.get("GITHUB_APP_CLIENT_ID", "")
|
|
|
|
|
GITHUB_APP_CLIENT_SECRET = os.environ.get("GITHUB_APP_CLIENT_SECRET", "")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def _secret() -> str:
|
|
|
|
|
s = os.environ.get("DASHBOARD_JWT_SECRET", "")
|
|
|
|
|
if not s:
|
|
|
|
|
raise HTTPException(500, "DASHBOARD_JWT_SECRET not configured")
|
|
|
|
|
return s
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def _allowed_redirect_origins() -> set[str]:
|
|
|
|
|
"""Origins permitted for the post-login redirect.
|
|
|
|
|
|
|
|
|
|
Built from DASHBOARD_BASE_URL plus any DASHBOARD_ALLOWED_ORIGINS entries
|
|
|
|
|
so the dashboard itself and its preview deploys can all be redirect
|
|
|
|
|
targets — but nothing else.
|
|
|
|
|
"""
|
|
|
|
|
origins: set[str] = set()
|
|
|
|
|
base = os.environ.get("DASHBOARD_BASE_URL", "").strip()
|
|
|
|
|
if base:
|
|
|
|
|
origins.add(_origin_of(base))
|
|
|
|
|
for entry in os.environ.get("DASHBOARD_ALLOWED_ORIGINS", "").split(","):
|
|
|
|
|
entry = entry.strip()
|
|
|
|
|
if entry:
|
|
|
|
|
origins.add(_origin_of(entry))
|
|
|
|
|
origins.discard("")
|
|
|
|
|
return origins
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def _origin_of(url: str) -> str:
|
|
|
|
|
parsed = urlparse(url)
|
|
|
|
|
if not parsed.scheme or not parsed.netloc:
|
|
|
|
|
return ""
|
|
|
|
|
return f"{parsed.scheme}://{parsed.netloc}"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def sanitize_redirect_to(redirect_to: str | None) -> str:
|
|
|
|
|
"""Return a safe post-login redirect URL.
|
|
|
|
|
|
|
|
|
|
Falls back to DASHBOARD_BASE_URL when the supplied URL's origin isn't
|
|
|
|
|
explicitly allowed. This blocks the open-redirect / phishing primitive
|
|
|
|
|
where an attacker drops their own URL into `?redirect_to=`.
|
|
|
|
|
"""
|
|
|
|
|
fallback = os.environ.get("DASHBOARD_BASE_URL", "").strip()
|
|
|
|
|
if not redirect_to:
|
|
|
|
|
return fallback
|
|
|
|
|
candidate_origin = _origin_of(redirect_to)
|
|
|
|
|
if not candidate_origin:
|
|
|
|
|
return fallback
|
|
|
|
|
if candidate_origin in _allowed_redirect_origins():
|
|
|
|
|
return redirect_to
|
|
|
|
|
logger.warning("Rejected redirect_to=%r — origin not in allowlist", redirect_to)
|
|
|
|
|
return fallback
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def issue_session(*, login: str, email: str | None, avatar_url: str | None) -> str:
|
|
|
|
|
now = int(time.time())
|
|
|
|
|
payload = {
|
|
|
|
|
"sub": login,
|
|
|
|
|
"email": email,
|
|
|
|
|
"avatar_url": avatar_url,
|
|
|
|
|
"iat": now,
|
|
|
|
|
"exp": now + SESSION_TTL_SECONDS,
|
|
|
|
|
}
|
|
|
|
|
return jwt.encode(payload, _secret(), algorithm=JWT_ALG)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def decode_session(token: str) -> dict[str, Any]:
|
|
|
|
|
try:
|
|
|
|
|
return jwt.decode(token, _secret(), algorithms=[JWT_ALG])
|
|
|
|
|
except jwt.PyJWTError as e:
|
|
|
|
|
raise HTTPException(401, f"invalid session: {e}") from e
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def new_state_nonce() -> str:
|
|
|
|
|
"""A fresh random nonce used to bind state JWT ↔ browser cookie."""
|
|
|
|
|
return secrets.token_urlsafe(32)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def hash_state_nonce(nonce: str) -> str:
|
|
|
|
|
"""HMAC the nonce so the value stored on the wire isn't reversible.
|
|
|
|
|
|
|
|
|
|
We compare ``hash_state_nonce(cookie_nonce) == state.nonce_hash`` at
|
|
|
|
|
callback time. Using HMAC over a constant-time digest also gives us
|
|
|
|
|
timing-attack resistance via :func:`hmac.compare_digest` at the call
|
|
|
|
|
site.
|
|
|
|
|
"""
|
|
|
|
|
return hmac.new(_secret().encode(), nonce.encode(), hashlib.sha256).hexdigest()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def issue_state(*, redirect_to: str, nonce_hash: str) -> str:
|
|
|
|
|
now = int(time.time())
|
|
|
|
|
payload = {
|
|
|
|
|
"nonce_hash": nonce_hash,
|
|
|
|
|
"redirect_to": redirect_to,
|
|
|
|
|
"iat": now,
|
|
|
|
|
"exp": now + STATE_TTL_SECONDS,
|
|
|
|
|
}
|
|
|
|
|
return jwt.encode(payload, _secret(), algorithm=JWT_ALG)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def decode_state(state: str) -> dict[str, Any]:
|
|
|
|
|
try:
|
|
|
|
|
return jwt.decode(state, _secret(), algorithms=[JWT_ALG])
|
|
|
|
|
except jwt.PyJWTError as e:
|
|
|
|
|
raise HTTPException(400, f"invalid state: {e}") from e
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
def require_session(request: Request) -> dict[str, Any]:
|
|
|
|
|
token = request.cookies.get(COOKIE_NAME)
|
|
|
|
|
if not token:
|
|
|
|
|
raise HTTPException(401, "not authenticated")
|
|
|
|
|
return decode_session(token)
|
|
|
|
|
|
|
|
|
|
|
2026-05-21 10:44:04 -07:00
|
|
|
def expires_at_from_github_response(data: dict[str, Any], *, field: str) -> str | None:
|
|
|
|
|
"""Convert GitHub ``expires_in`` / ``refresh_token_expires_in`` to an ISO timestamp."""
|
|
|
|
|
raw = data.get(field)
|
|
|
|
|
if not isinstance(raw, int | float) or raw <= 0:
|
|
|
|
|
return None
|
|
|
|
|
return (datetime.now(UTC) + timedelta(seconds=int(raw))).isoformat()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
async def _request_github_tokens(body: dict[str, str]) -> dict[str, Any]:
|
feat: open-swe dashboard for per-user profile config (#1302)
* feat: dashboard backend — GitHub OAuth, profile CRUD, admin endpoints
Adds agent/dashboard/ FastAPI router mounted at /dashboard/api covering:
- GitHub App OAuth login → JWT cookie session (cross-domain ready)
- profile CRUD against LangGraph Store with model+effort validation
- admin gate via CONFIGURED_ADMINS
- /repos via /user/installations using the user's encrypted OAuth token
CORS allowlist on webapp.py is opt-in via DASHBOARD_ALLOWED_ORIGINS so the
Vercel-hosted frontend can call the LangSmith deployment with credentials.
* feat: apply dashboard profile model/effort overrides in get_agent
Look up the triggering user's GitHub login from config (direct field or
GITHUB_USER_EMAIL_MAP reverse lookup), read their profile from the Store,
and apply default_model + reasoning_effort to make_model when both are
valid. Effort 'max' is captured on the profile but not yet wired through —
the OpenAI Reasoning Literal doesn't accept it.
* feat: ui/ TanStack Start dashboard for profile config
Scaffolded with the shadcn b7CScJIjA preset (TanStack Start template,
base-ui primitives, Tailwind v4). Three routes:
- /login — Sign in with GitHub (links to /dashboard/api/auth/login)
- /profile — Edit default model, reasoning effort, default repo
- /admin — Admin-only: list users and edit other profiles
API client (src/lib/api.ts) uses credentials: include so the osw_session
cookie set by the OAuth callback rides cross-origin. VITE_DASHBOARD_API_BASE_URL
points at the LangSmith deployment.
Effort options re-render when the model changes; 'max' on Opus 4.7 is
captured on the profile but ignored downstream until anthropic reasoning
is wired through make_model.
* feat: searchable Combobox for default repo picker
Replaces the Select with a base-ui Combobox so users can filter by typing,
the popup is wider than the trigger so full owner/repo names are readable,
and the list caps at max-h-80 to stay on screen.
* fix: address review comments + wire default_repo and Anthropic thinking
Security/correctness fixes from PR review:
* Open redirect: validate `redirect_to` in `/auth/login` against
`DASHBOARD_BASE_URL` + `DASHBOARD_ALLOWED_ORIGINS` before signing it
into the state JWT. Anything off-allowlist falls back to the dashboard
base URL. (PR #1302 r3250054386)
* Login CSRF: bind the OAuth `state` to the requesting browser. At
`/auth/login` we generate a fresh nonce, set it as a short-lived
HttpOnly SameSite=Lax cookie scoped to `/dashboard/api/auth`, and
embed `hash_state_nonce(nonce)` in the state JWT. At `/auth/callback`
we require the cookie nonce to hash-match the state JWT's nonce_hash
(constant-time compare). (PR #1302 r3250054395)
* RMW race in profile vs token writes: split storage into two
namespaces — `["profiles"]` for user-editable settings and
`["oauth_tokens"]` for the encrypted GitHub token. Each upsert now
only writes its own namespace so an in-flight profile save can no
longer clobber a fresh token from a concurrent re-login (and vice
versa). (PR #1302 r3250054393)
* /repos pagination: follow `Link: rel="next"` for both
`/user/installations` and per-installation `/repositories` with
per_page=100, capped at 1000 items. (PR #1302 r3250054401)
Feature wires:
* default_repo: applied as a fallback in `get_slack_repo_config` (after
explicit-repo / thread metadata, before the env defaults) and in the
Linear webhook (after comment-body extraction, before team mapping).
Both paths resolve the triggering user's GitHub login via
GITHUB_USER_EMAIL_MAP and read the profile's default_repo.
* Anthropic "thinking" effort: `make_model` now accepts a `thinking`
kwarg; `get_agent` maps profile effort {low,medium,high,xhigh,max}
to budget_tokens {1k,4k,12k,32k,60k} when the chosen model is
anthropic. OpenAI path still ignores "max" since the Literal doesn't
accept it.
2026-05-15 11:23:53 -07:00
|
|
|
if not GITHUB_APP_CLIENT_ID or not GITHUB_APP_CLIENT_SECRET:
|
|
|
|
|
raise HTTPException(500, "GitHub App OAuth not configured")
|
|
|
|
|
async with httpx.AsyncClient() as client:
|
|
|
|
|
resp = await client.post(
|
|
|
|
|
"https://github.com/login/oauth/access_token",
|
|
|
|
|
headers={"Accept": "application/json"},
|
2026-05-21 10:44:04 -07:00
|
|
|
data=body,
|
feat: open-swe dashboard for per-user profile config (#1302)
* feat: dashboard backend — GitHub OAuth, profile CRUD, admin endpoints
Adds agent/dashboard/ FastAPI router mounted at /dashboard/api covering:
- GitHub App OAuth login → JWT cookie session (cross-domain ready)
- profile CRUD against LangGraph Store with model+effort validation
- admin gate via CONFIGURED_ADMINS
- /repos via /user/installations using the user's encrypted OAuth token
CORS allowlist on webapp.py is opt-in via DASHBOARD_ALLOWED_ORIGINS so the
Vercel-hosted frontend can call the LangSmith deployment with credentials.
* feat: apply dashboard profile model/effort overrides in get_agent
Look up the triggering user's GitHub login from config (direct field or
GITHUB_USER_EMAIL_MAP reverse lookup), read their profile from the Store,
and apply default_model + reasoning_effort to make_model when both are
valid. Effort 'max' is captured on the profile but not yet wired through —
the OpenAI Reasoning Literal doesn't accept it.
* feat: ui/ TanStack Start dashboard for profile config
Scaffolded with the shadcn b7CScJIjA preset (TanStack Start template,
base-ui primitives, Tailwind v4). Three routes:
- /login — Sign in with GitHub (links to /dashboard/api/auth/login)
- /profile — Edit default model, reasoning effort, default repo
- /admin — Admin-only: list users and edit other profiles
API client (src/lib/api.ts) uses credentials: include so the osw_session
cookie set by the OAuth callback rides cross-origin. VITE_DASHBOARD_API_BASE_URL
points at the LangSmith deployment.
Effort options re-render when the model changes; 'max' on Opus 4.7 is
captured on the profile but ignored downstream until anthropic reasoning
is wired through make_model.
* feat: searchable Combobox for default repo picker
Replaces the Select with a base-ui Combobox so users can filter by typing,
the popup is wider than the trigger so full owner/repo names are readable,
and the list caps at max-h-80 to stay on screen.
* fix: address review comments + wire default_repo and Anthropic thinking
Security/correctness fixes from PR review:
* Open redirect: validate `redirect_to` in `/auth/login` against
`DASHBOARD_BASE_URL` + `DASHBOARD_ALLOWED_ORIGINS` before signing it
into the state JWT. Anything off-allowlist falls back to the dashboard
base URL. (PR #1302 r3250054386)
* Login CSRF: bind the OAuth `state` to the requesting browser. At
`/auth/login` we generate a fresh nonce, set it as a short-lived
HttpOnly SameSite=Lax cookie scoped to `/dashboard/api/auth`, and
embed `hash_state_nonce(nonce)` in the state JWT. At `/auth/callback`
we require the cookie nonce to hash-match the state JWT's nonce_hash
(constant-time compare). (PR #1302 r3250054395)
* RMW race in profile vs token writes: split storage into two
namespaces — `["profiles"]` for user-editable settings and
`["oauth_tokens"]` for the encrypted GitHub token. Each upsert now
only writes its own namespace so an in-flight profile save can no
longer clobber a fresh token from a concurrent re-login (and vice
versa). (PR #1302 r3250054393)
* /repos pagination: follow `Link: rel="next"` for both
`/user/installations` and per-installation `/repositories` with
per_page=100, capped at 1000 items. (PR #1302 r3250054401)
Feature wires:
* default_repo: applied as a fallback in `get_slack_repo_config` (after
explicit-repo / thread metadata, before the env defaults) and in the
Linear webhook (after comment-body extraction, before team mapping).
Both paths resolve the triggering user's GitHub login via
GITHUB_USER_EMAIL_MAP and read the profile's default_repo.
* Anthropic "thinking" effort: `make_model` now accepts a `thinking`
kwarg; `get_agent` maps profile effort {low,medium,high,xhigh,max}
to budget_tokens {1k,4k,12k,32k,60k} when the chosen model is
anthropic. OpenAI path still ignores "max" since the Literal doesn't
accept it.
2026-05-15 11:23:53 -07:00
|
|
|
)
|
|
|
|
|
resp.raise_for_status()
|
|
|
|
|
data = resp.json()
|
2026-05-21 10:44:04 -07:00
|
|
|
if not isinstance(data, dict):
|
|
|
|
|
raise HTTPException(502, "unexpected GitHub OAuth response")
|
|
|
|
|
if data.get("error"):
|
|
|
|
|
raise HTTPException(
|
|
|
|
|
400, f"github oauth error: {data.get('error_description') or data['error']}"
|
|
|
|
|
)
|
|
|
|
|
return data
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
async def exchange_code(code: str) -> dict[str, Any]:
|
|
|
|
|
"""Exchange an OAuth authorization code for user-to-server tokens."""
|
|
|
|
|
data = await _request_github_tokens(
|
|
|
|
|
{
|
|
|
|
|
"client_id": GITHUB_APP_CLIENT_ID,
|
|
|
|
|
"client_secret": GITHUB_APP_CLIENT_SECRET,
|
|
|
|
|
"code": code,
|
|
|
|
|
}
|
|
|
|
|
)
|
|
|
|
|
if not data.get("access_token"):
|
feat: open-swe dashboard for per-user profile config (#1302)
* feat: dashboard backend — GitHub OAuth, profile CRUD, admin endpoints
Adds agent/dashboard/ FastAPI router mounted at /dashboard/api covering:
- GitHub App OAuth login → JWT cookie session (cross-domain ready)
- profile CRUD against LangGraph Store with model+effort validation
- admin gate via CONFIGURED_ADMINS
- /repos via /user/installations using the user's encrypted OAuth token
CORS allowlist on webapp.py is opt-in via DASHBOARD_ALLOWED_ORIGINS so the
Vercel-hosted frontend can call the LangSmith deployment with credentials.
* feat: apply dashboard profile model/effort overrides in get_agent
Look up the triggering user's GitHub login from config (direct field or
GITHUB_USER_EMAIL_MAP reverse lookup), read their profile from the Store,
and apply default_model + reasoning_effort to make_model when both are
valid. Effort 'max' is captured on the profile but not yet wired through —
the OpenAI Reasoning Literal doesn't accept it.
* feat: ui/ TanStack Start dashboard for profile config
Scaffolded with the shadcn b7CScJIjA preset (TanStack Start template,
base-ui primitives, Tailwind v4). Three routes:
- /login — Sign in with GitHub (links to /dashboard/api/auth/login)
- /profile — Edit default model, reasoning effort, default repo
- /admin — Admin-only: list users and edit other profiles
API client (src/lib/api.ts) uses credentials: include so the osw_session
cookie set by the OAuth callback rides cross-origin. VITE_DASHBOARD_API_BASE_URL
points at the LangSmith deployment.
Effort options re-render when the model changes; 'max' on Opus 4.7 is
captured on the profile but ignored downstream until anthropic reasoning
is wired through make_model.
* feat: searchable Combobox for default repo picker
Replaces the Select with a base-ui Combobox so users can filter by typing,
the popup is wider than the trigger so full owner/repo names are readable,
and the list caps at max-h-80 to stay on screen.
* fix: address review comments + wire default_repo and Anthropic thinking
Security/correctness fixes from PR review:
* Open redirect: validate `redirect_to` in `/auth/login` against
`DASHBOARD_BASE_URL` + `DASHBOARD_ALLOWED_ORIGINS` before signing it
into the state JWT. Anything off-allowlist falls back to the dashboard
base URL. (PR #1302 r3250054386)
* Login CSRF: bind the OAuth `state` to the requesting browser. At
`/auth/login` we generate a fresh nonce, set it as a short-lived
HttpOnly SameSite=Lax cookie scoped to `/dashboard/api/auth`, and
embed `hash_state_nonce(nonce)` in the state JWT. At `/auth/callback`
we require the cookie nonce to hash-match the state JWT's nonce_hash
(constant-time compare). (PR #1302 r3250054395)
* RMW race in profile vs token writes: split storage into two
namespaces — `["profiles"]` for user-editable settings and
`["oauth_tokens"]` for the encrypted GitHub token. Each upsert now
only writes its own namespace so an in-flight profile save can no
longer clobber a fresh token from a concurrent re-login (and vice
versa). (PR #1302 r3250054393)
* /repos pagination: follow `Link: rel="next"` for both
`/user/installations` and per-installation `/repositories` with
per_page=100, capped at 1000 items. (PR #1302 r3250054401)
Feature wires:
* default_repo: applied as a fallback in `get_slack_repo_config` (after
explicit-repo / thread metadata, before the env defaults) and in the
Linear webhook (after comment-body extraction, before team mapping).
Both paths resolve the triggering user's GitHub login via
GITHUB_USER_EMAIL_MAP and read the profile's default_repo.
* Anthropic "thinking" effort: `make_model` now accepts a `thinking`
kwarg; `get_agent` maps profile effort {low,medium,high,xhigh,max}
to budget_tokens {1k,4k,12k,32k,60k} when the chosen model is
anthropic. OpenAI path still ignores "max" since the Literal doesn't
accept it.
2026-05-15 11:23:53 -07:00
|
|
|
raise HTTPException(400, f"oauth exchange failed: {data}")
|
2026-05-21 10:44:04 -07:00
|
|
|
return data
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
async def refresh_user_access_token(refresh_token: str) -> dict[str, Any]:
|
|
|
|
|
"""Rotate an expiring user access token using its refresh token."""
|
|
|
|
|
data = await _request_github_tokens(
|
|
|
|
|
{
|
|
|
|
|
"client_id": GITHUB_APP_CLIENT_ID,
|
|
|
|
|
"client_secret": GITHUB_APP_CLIENT_SECRET,
|
|
|
|
|
"grant_type": "refresh_token",
|
|
|
|
|
"refresh_token": refresh_token,
|
|
|
|
|
}
|
|
|
|
|
)
|
|
|
|
|
if not data.get("access_token"):
|
|
|
|
|
raise HTTPException(400, f"oauth refresh failed: {data}")
|
|
|
|
|
return data
|
feat: open-swe dashboard for per-user profile config (#1302)
* feat: dashboard backend — GitHub OAuth, profile CRUD, admin endpoints
Adds agent/dashboard/ FastAPI router mounted at /dashboard/api covering:
- GitHub App OAuth login → JWT cookie session (cross-domain ready)
- profile CRUD against LangGraph Store with model+effort validation
- admin gate via CONFIGURED_ADMINS
- /repos via /user/installations using the user's encrypted OAuth token
CORS allowlist on webapp.py is opt-in via DASHBOARD_ALLOWED_ORIGINS so the
Vercel-hosted frontend can call the LangSmith deployment with credentials.
* feat: apply dashboard profile model/effort overrides in get_agent
Look up the triggering user's GitHub login from config (direct field or
GITHUB_USER_EMAIL_MAP reverse lookup), read their profile from the Store,
and apply default_model + reasoning_effort to make_model when both are
valid. Effort 'max' is captured on the profile but not yet wired through —
the OpenAI Reasoning Literal doesn't accept it.
* feat: ui/ TanStack Start dashboard for profile config
Scaffolded with the shadcn b7CScJIjA preset (TanStack Start template,
base-ui primitives, Tailwind v4). Three routes:
- /login — Sign in with GitHub (links to /dashboard/api/auth/login)
- /profile — Edit default model, reasoning effort, default repo
- /admin — Admin-only: list users and edit other profiles
API client (src/lib/api.ts) uses credentials: include so the osw_session
cookie set by the OAuth callback rides cross-origin. VITE_DASHBOARD_API_BASE_URL
points at the LangSmith deployment.
Effort options re-render when the model changes; 'max' on Opus 4.7 is
captured on the profile but ignored downstream until anthropic reasoning
is wired through make_model.
* feat: searchable Combobox for default repo picker
Replaces the Select with a base-ui Combobox so users can filter by typing,
the popup is wider than the trigger so full owner/repo names are readable,
and the list caps at max-h-80 to stay on screen.
* fix: address review comments + wire default_repo and Anthropic thinking
Security/correctness fixes from PR review:
* Open redirect: validate `redirect_to` in `/auth/login` against
`DASHBOARD_BASE_URL` + `DASHBOARD_ALLOWED_ORIGINS` before signing it
into the state JWT. Anything off-allowlist falls back to the dashboard
base URL. (PR #1302 r3250054386)
* Login CSRF: bind the OAuth `state` to the requesting browser. At
`/auth/login` we generate a fresh nonce, set it as a short-lived
HttpOnly SameSite=Lax cookie scoped to `/dashboard/api/auth`, and
embed `hash_state_nonce(nonce)` in the state JWT. At `/auth/callback`
we require the cookie nonce to hash-match the state JWT's nonce_hash
(constant-time compare). (PR #1302 r3250054395)
* RMW race in profile vs token writes: split storage into two
namespaces — `["profiles"]` for user-editable settings and
`["oauth_tokens"]` for the encrypted GitHub token. Each upsert now
only writes its own namespace so an in-flight profile save can no
longer clobber a fresh token from a concurrent re-login (and vice
versa). (PR #1302 r3250054393)
* /repos pagination: follow `Link: rel="next"` for both
`/user/installations` and per-installation `/repositories` with
per_page=100, capped at 1000 items. (PR #1302 r3250054401)
Feature wires:
* default_repo: applied as a fallback in `get_slack_repo_config` (after
explicit-repo / thread metadata, before the env defaults) and in the
Linear webhook (after comment-body extraction, before team mapping).
Both paths resolve the triggering user's GitHub login via
GITHUB_USER_EMAIL_MAP and read the profile's default_repo.
* Anthropic "thinking" effort: `make_model` now accepts a `thinking`
kwarg; `get_agent` maps profile effort {low,medium,high,xhigh,max}
to budget_tokens {1k,4k,12k,32k,60k} when the chosen model is
anthropic. OpenAI path still ignores "max" since the Literal doesn't
accept it.
2026-05-15 11:23:53 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
async def fetch_github_user(access_token: str) -> tuple[dict[str, Any], str | None]:
|
|
|
|
|
"""Return ``(user, primary_email)`` for the authenticated user."""
|
|
|
|
|
headers = {
|
|
|
|
|
"Authorization": f"Bearer {access_token}",
|
|
|
|
|
"Accept": "application/vnd.github+json",
|
|
|
|
|
"X-GitHub-Api-Version": "2022-11-28",
|
|
|
|
|
}
|
|
|
|
|
async with httpx.AsyncClient() as client:
|
|
|
|
|
u = await client.get("https://api.github.com/user", headers=headers)
|
|
|
|
|
u.raise_for_status()
|
|
|
|
|
user = u.json()
|
|
|
|
|
email = user.get("email")
|
|
|
|
|
if not email:
|
|
|
|
|
e = await client.get("https://api.github.com/user/emails", headers=headers)
|
|
|
|
|
if e.status_code == 200:
|
|
|
|
|
primary = next((x for x in e.json() if x.get("primary")), None)
|
|
|
|
|
if primary:
|
|
|
|
|
email = primary.get("email")
|
|
|
|
|
return user, email
|