- use tsx for bl/kscli dev and test runners
- point local library exports to src while keeping publishConfig on dist
- switch vendored telemetry modules to .cjs for source-run compatibility
- update stress/e2e helpers and agent docs for the new source execution path
- add openapi auth requirement with command-scoped access key flags and paired credential resolution
- persist OpenAPI credentials through auth login/status/logout using access_key_* config fields
- route token-plan commands through the centralized ACS signing client
- keep legacy openapi_access_key_* config readable while rejecting it as a new config set key
- refresh docs, generated references, telemetry authMethod, and e2e coverage
- remove --no-color from GLOBAL_FLAGS and drop Settings.noColor
- move ANSI styling decisions into runtime color helpers with NO_COLOR support
- update command text renderers to use shared color helpers instead of local ANSI codes
- refresh e2e invocations, generated reference, and agent skill guidance
- scope image/video task execution to --async and --concurrent
- add concurrent task fan-out for video edit and video ref
- extend image edit to the async image task path
- refresh e2e coverage and generated command references
- remove nonInteractive plus yes/async/concurrent from GLOBAL_FLAGS and Settings;
command dispatch no longer resolves command-only switches into global settings
- add shared ASYNC_FLAG / CONCURRENT_FLAG definitions for commands that actually
support task-only return or parallel requests
- keep quota downgrade protection by moving --yes onto quota request and reading
flags.yes for confirmed downgrade submission
- update existing async/concurrent consumers to read own flags; no new capability
matrix entries are added
- refresh generated command reference and remove stale --non-interactive usage
from e2e/stress invocations
- flags split into GLOBAL_FLAGS (all commands) plus MODEL_AUTH_FLAGS /
CONSOLE_AUTH_FLAGS, parsed only for commands of the matching auth
domain; cross-domain flags now fail with "Unknown flag" instead of
being silently ignored
- all shadow redeclarations removed; the registry guard now rejects any
own flag named after a reserved (global or visible-domain) flag
- --workspace-id joins the console domain (chain: flag > env > file);
usage stats drops its private declaration and in-command priority
- auth login declares its credential args as own command parameters
(--api-key / --base-url / --console-site, original behavior intact);
auth status no longer accepts credential-domain overrides (use env or
config set instead)
- command help and the generated reference both show Flags (own + auth
domain) plus a full Global Flags section, replacing the footer hint
- breaking: pipeline run --timeout renamed to --step-timeout (collided
with the global request timeout)
Missing required flags now throw UsageError (exit code 2, error on
stderr, JSON under --output json) instead of printing help and exiting
0. Update assertions and titles accordingly:
- missing-flag cases: exitCode 0/1 -> 2, retitle to "errors as usage error (2)"
- auth / video-task-get under --output json: assert the error JSON on stderr
- proxy probe: import setupProxyFromEnv from runtime/src/proxy.ts
- drop tests for the removed config export-schema command
- fix t2v dry-run: cliTimeoutPrefix misplaced between --model and its value
Commands no longer call process.exit() directly. Every failure now throws
UsageError (bad input → exit 2) or BailianError (runtime failure, with
AUTH/TIMEOUT codes), so the runtime's handleError stays the single exit
point and telemetry always flushes.
- convert 27 process.exit() sites across 15 commands to throws
- move cross-flag/value checks into validate(); use flag `choices` for
--events / --sort; drop dead --model required check
- keep pipeline's process.exitCode for lint-style soft failures
- enforce with unicorn/no-process-exit, allowed only in runtime/tools/tests
- fix incidental lint: void floating run(), narrow console errorCode,
align generate-reference type imports to source
AK/SK signing was used only by `knowledge retrieve`'s deprecated fallback,
which the api-key auth gate now makes unreachable. Drop it; the command is
pure api-key.
- knowledge/retrieve: remove the AK/SK path + --access-key-id/secret/workspace-id
flags; api-key only
- delete client/ak-sign.ts and its signRequest/AkSignConfig exports
- drop access_key_id/access_key_secret from config schema, loader, and
`config show` / `config set`
- remove the now-unused PascalCase KnowledgeRetrieve request/response types
Move all credential handling out of commands into one place. authStage
resolves the credential for the command's declared `auth` and bakes it into
`ctx.client`, gating (throw) when missing; commands reach the network only
through `ctx.client` and never touch tokens or baseUrl.
- add Client (request/requestJson/uploadFile/mcp/console/url) wrapping the
token + baseUrl; commands call it instead of self-resolving
- run(config, flags) → run(ctx) + CommandContext; migrate all 45 commands
- split domains: model = pure api-key (drop access-token fallback), console =
config.json only (drop DASHSCOPE_ACCESS_TOKEN env)
- consolidate env reads in loadConfig; CredentialSource = flag | env | config;
priority flag > env > config
- endpoints return paths (xxxPath) instead of full URLs; baseUrl owned by Client
- auth status now uses describeAuth
- video/download: auth "none" → "apiKey" (it needs the model API)
- dedupe fetchModelList behind an injected console-call function
- remove ensureApiKey/ensure-key.ts, prompt.ts + isInteractive, and the old
resolveCredential/resolveConsoleGatewayCredential resolvers
- tests: adapt auth.e2e to the new auth status shape; drop the
DASHSCOPE_ACCESS_TOKEN branch from console-readiness gates
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
若麒