Eliminate per-command --region/--site/--switch-agent duplication across 11 console gateway commands.
These values now flow through config (CLI flags → config file → defaults) and are consumed by
callConsoleGateway automatically. Also wire consoleSite into resolveConsoleOrigin so --console-site
selects the correct login URL (domestic vs international).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Node's built-in fetch (undici) ignores proxy environment variables, so
bl always connected directly and failed with ECONNRESET behind a VPN or
corporate proxy. Install an EnvHttpProxyAgent as the global dispatcher at
startup, but only when a proxy variable is actually set — behavior is
unchanged otherwise. Lowercase variables take precedence over uppercase
(curl convention) and NO_PROXY is honored.
Values are trimmed and passed explicitly to work around undici reading
env vars with ??, where an empty lowercase variable (https_proxy="")
masks a configured uppercase one. Invalid proxy URLs fail with a clear
usage error instead of a stack trace, and the ECONNRESET hint now
suggests exporting HTTPS_PROXY.
Tests are fully offline and need no credentials: unit tests cover env
parsing, and the e2e test runs a minimal probe (setupProxyFromEnv + a
bare fetch) against a .invalid host through a local CONNECT proxy to
verify traffic routes through the proxy, NO_PROXY is honored, no
dispatcher is installed when unset, and invalid values error clearly.
- `bl mcp list` — list MCP servers enabled under the current Bailian
account via console gateway PageList (always activated=1).
- `bl mcp tools <server-code>` — list tools exposed by a server.
- `bl mcp call <server-code>.<tool>` — invoke a tool. Accepts `--json`,
repeatable `--arg k=v` (JSON-parsed when possible) and `--query` sugar;
`--url` overrides the endpoint for non-Bailian MCPs.
- core: export `bailianMcpUrl(baseUrl, code)` building
`/api/v1/mcps/<code>/mcp`; `McpClient` now takes a full URL.
- `bl search web` switches to `mcpWebSearchEndpoint` directly.
- e2e: `mcp.e2e.test.ts` covering help, dry-run, arg-merge semantics,
invalid-input paths, and one live `tools/list` against WebSearch.
script/js executes its `code` as host JavaScript (via new Function), and a step's
`code` is a *resolved* input — so it could be written as `{ $from: <chat-step> }`,
turning model/API output into the body of the executed function (untrusted data
-> arbitrary host code execution). Pipeline validation now requires script/js
`code` to be a literal string: any $from/expression-sourced code is rejected.
Authoring a literal script/js step remains supported (the pipeline file is the
trust boundary, like a shell/npm script). Combined with "dry-run never executes
$js", this closes the path where untrusted text reaches the JS sink.
Adds regression tests: $from-sourced code rejected, literal code accepted,
dry-run does not execute $js, getByJsonPointer blocks prototype/inherited keys,
and concurrency clamps to the maximum.
https://claude.ai/code/session_017ZGQCjwNQF5Pz96gLUnnG1
ls