feat(auth): add PAT auth, scoped REST/MCP access, CSRF, and tenant isolation

Introduce Personal Access Tokens (nzp_*) as a stateless auth path alongside
JWT, gated per-endpoint by a scope middleware (nezha:{resource}:{verb}) with
fail-closed empty-scope defaults and a server-id whitelist. Self-management
endpoints (profile, api-tokens, oauth2 bind, refresh-token) explicitly reject
PATs to block privilege-escalation chains. A revoke registry tears down active
long-lived connections (terminal, fm, ws, transfer, mcp) the moment a PAT is
deleted, with a tombstone closing the revoke->register race.

Add an MCP endpoint that proxies tool calls (exec, fs read/write/delete,
transfer) to agents over gRPC, guarded by origin/DNS-rebinding checks, a
per-token rate limiter, audit logging, and a kill switch. Serialize all
sends through the IOStream wrapper to honour grpc-go's concurrency contract.

Add CSRF double-submit protection on unsafe cookie-authenticated methods,
exempting authenticated PAT requests by context identity (not a forgeable
Authorization header). Apply visibility/whitelist filtering consistently
across list, get-by-id, and mutate paths to enforce tenant isolation.

Migrate legacy mcp:* scopes: rewrite read/exec to nezha:* equivalents and
drop dangerous write/delete/wildcard grants.

Co-authored-by: cloudcode <cloudcode@users.noreply.github.com>
This commit is contained in:
naiba
2026-05-30 15:56:44 +00:00
co-authored by cloudcode
parent 58c1ea7b0b
commit ab25662ddd
153 changed files with 16974 additions and 244 deletions
@@ -0,0 +1,63 @@
package controller
import (
"encoding/json"
"testing"
"github.com/stretchr/testify/require"
"github.com/nezhahq/nezha/model"
)
func TestServerExec_RejectsOutOfRangeTimeoutSeconds(t *testing.T) {
cleanup, uid := setupMCPTest(t)
defer cleanup()
tok, _ := mkToken(t, uid, []string{model.ScopeServerExec}, nil)
// timeout_seconds is documented as 1..300 in the tool schema; sending
// 1_000_000 would otherwise let the dashboard wait ~1e6s on rpc.CallAgent
// when the agent is unreachable / old, turning one MCP call into a
// long-lived goroutine + connection occupation. Handler must reject
// before touching the RPC layer.
c, w := mcpCallCtx(t, tok, uid, jsonRPCRequest{
JSONRPC: "2.0", ID: json.RawMessage("1"), Method: "tools/call",
Params: jsonObj(t, toolCallParams{
Name: "server.exec",
Arguments: jsonRaw(map[string]any{
"server_id": 7,
"cmd": "echo",
"timeout_seconds": 1_000_000,
}),
}),
})
mcpEndpoint(c)
_, tcr := decodeRPC(w)
require.NotNil(t, tcr)
require.True(t, tcr.IsError, "expected isError=true for out-of-range timeout, got %+v", tcr)
require.Contains(t, tcr.Content[0].Text, "timeout_seconds")
}
func TestServerExec_RejectsZeroLikeNegativeTimeoutBoundary(t *testing.T) {
cleanup, uid := setupMCPTest(t)
defer cleanup()
tok, _ := mkToken(t, uid, []string{model.ScopeServerExec}, nil)
// 301s sits one above the documented maximum. The previous handler
// happily forwarded it as-is and added +5s to the dashboard-side wait,
// so any client could ignore the schema bound. Pin the rejection.
c, w := mcpCallCtx(t, tok, uid, jsonRPCRequest{
JSONRPC: "2.0", ID: json.RawMessage("1"), Method: "tools/call",
Params: jsonObj(t, toolCallParams{
Name: "server.exec",
Arguments: jsonRaw(map[string]any{
"server_id": 7,
"cmd": "echo",
"timeout_seconds": 301,
}),
}),
})
mcpEndpoint(c)
_, tcr := decodeRPC(w)
require.True(t, tcr.IsError)
require.Contains(t, tcr.Content[0].Text, "timeout_seconds")
}