feat(channels/yuanbao): add Yuanbao channel provider

[senamakel]

Summary

• New Yuanbao channel provider (Tencent 元宝 robot logic v5) under src/openhuman/channels/providers/yuanbao/, wired into the channel registry + startup loop and exposed to the UI through the existing ChannelSetupModal flow.
• Outbound: HMAC-SHA256 sign-token client (sign.rs) with cached tokens, message splitting (splitter.rs), and a COS uploader (cos.rs) for media.
• Inbound: framed protobuf WebSocket client (connection.rs + wire.rs + proto_biz.rs) with bounded reconnect, single-flight token refresh on code=10099, prompt shutdown response, and bounded length-delimited parsing.
• Frontend: YuanbaoConfig.tsx form (i18n-keyed validation messages), YuanbaoIcon, and ChannelSetupModal integration for icon parity with ChannelSelector.
• Backend storage hygiene: app_secret is verified at connect time via the live sign-token endpoint, persisted only in the encrypted credentials store (never in plaintext config.toml), and hydrated at startup with a stale-app_key guard.

Problem

Yuanbao was not a supported channel; users on Tencent Yuanbao bots had no first-class way to wire OpenHuman to their robot.

Solution

This PR supersedes #2494 (from YuanbaoTeam/openhuman) because that PR did not have "Allow edits from maintainers" enabled and we could not push fixes directly. This version includes the original implementation plus:

1. Merge conflict resolution with current main (25 i18n chunk files).
2. Security fix: replaced 14 unconditional console.log/warn/error calls in YuanbaoConfig.tsx that leaked credential key names in production devtools — now uses the namespaced debug('channels:yuanbao') logger.
3. i18n fix: dedicated channels.yuanbao.{connect,reconnect,savedRestartRequired} keys instead of borrowing from channels.telegram.*.
4. Stability fix: replaced expect() panic in ops.rs with proper error propagation via ok_or_else()?.
5. Correctness fix: use definition.id instead of definition.icon for Yuanbao detection in ChannelSetupModal.
6. Build fix: fixed duplicate keys in de-5.ts causing TS1117 compile errors.
7. CodeRabbit fixes: re-entrancy guard, graceful shutdown cleanup, reconnect budget reset, varint overflow rejection.

Submission Checklist

If a section does not apply to this change, mark the item as N/A with a one-line reason. Do not delete items.

• Tests added or updated (happy path + at least one failure / edge case) per Testing Strategy
• Diff coverage ≥ 80% — changed lines (Vitest + cargo-llvm-cov merged via diff-cover) meet the gate enforced by .github/workflows/coverage.yml. Run pnpm test:coverage and pnpm test:rust locally; PRs below 80% on changed lines will not merge.
• Coverage matrix updated — added/removed/renamed feature rows in docs/TEST-COVERAGE-MATRIX.md reflect this change (or N/A: behaviour-only change)
• All affected feature IDs from the matrix are listed in the PR description under ## Related
• No new external network dependencies introduced (mock backend used per Testing Strategy)
• N/A: this PR does not touch release-cut surfaces (installer / code signing / OS permission prompts / DMG / Gatekeeper / accessibility)
• N/A: no prior GitHub issue; this PR introduces the channel directly. Supersedes feat(channels/yuanbao): add Yuanbao channel provider #2494.

Impact

• Desktop only. No Tauri shell changes; the channel runs entirely in the in-process Rust core. Added Rust deps were already present in the repo (reqwest, tungstenite, hmac/sha2). No web / mobile / CLI behavior changes outside the channel surface.
• Security: app_secret never lands in plaintext config.toml; runtime loads it from the encrypted credentials store at startup. Stale-app_key profiles will no longer be paired with the new key.
• Performance: sign-token cache + single-flight refresh; bounded WS backoff with prompt-exit on shutdown. No new periodic timers beyond the existing channel supervisor.

Related

• Supersedes: feat(channels/yuanbao): add Yuanbao channel provider #2494
• Matrix feature IDs: 10.1.5 Yuanbao Connection (added; layer RU, status 🟡)
• Follow-up PR(s)/TODOs: WDIO E2E spec for Yuanbao connect-flow, translate placeholder i18n strings in non-en/non-zh-CN locales.

Summary by CodeRabbit

Release Notes

• New Features

  • Added Yuanbao channel integration with full setup and configuration support.
  • New dedicated channel icon and UI components for Yuanbao.
  • Added multilingual support for Yuanbao across 15 languages.

• Tests

  • Added comprehensive test coverage for channel configuration and operations.

[coderabbitai]

Caution

Review failed

Failed to post review comments

📝 Walkthrough

Walkthrough

Adds a complete Yuanbao channel provider: frontend UI (icons, ChannelSetupModal, YuanbaoConfig), Redux/store updates, i18n strings, extensive Rust backend provider (config, SignManager, ops verification, WebSocket connection, inbound pipeline, outbound sender), protocol codecs, media/COS upload support, tests, and module wiring.

Changes

Yuanbao Channel Integration

Layer / File(s)	Summary
Channel type and backend definition src/openhuman/channels/controllers/definitions.rs, app/src/types/channels.ts	Register yuanbao channel definition and add 'yuanbao' to ChannelType.
Frontend icons & selector app/src/components/channels/YuanbaoIcon.tsx, app/src/components/channels/ChannelSelector.tsx, app/src/components/skills/skillIcons.tsx	Add YuanbaoIcon SVG component, use helper to render yuanbao icon, and expose in skill icons.
Channel setup UI app/src/components/channels/ChannelSetupModal.tsx, app/src/components/channels/YuanbaoConfig.tsx	Integrate YuanbaoConfig into modal; implement config UI with field validation, connect/disconnect flows, restart handling, and status badges.
Frontend tests app/src/components/channels/__tests__/*	Add Vitest+RTL tests for ChannelSetupModal routing, YuanbaoConfig workflows, and YuanbaoIcon uniqueness/DOM id handling.
Store & i18n app/src/store/channelConnectionsSlice.ts, app/src/lib/i18n/chunks/*, app/src/lib/i18n/en.ts	Defensive per-channel bucket initialization, migration additions for new channels, and add Yuanbao UI strings across language chunks.
Backend config & startup src/openhuman/channels/providers/yuanbao/config.rs, src/openhuman/config/schema/channels.rs, src/openhuman/channels/runtime/startup.rs	Add YuanbaoConfig struct, defaults/validation, wire startup to hydrate app_secret from credentials store and initialize channel.
Credential verification & ops tests src/openhuman/channels/providers/yuanbao/sign.rs, src/openhuman/channels/controllers/ops.rs, src/openhuman/channels/controllers/ops_tests.rs	Implement SignManager (HMAC-SHA256 signing + caching/coalescing), verify sign-token during connect_channel, persist config without plaintext secret, add wiremock tests.
Connection transport src/openhuman/channels/providers/yuanbao/connection.rs	Implement YuanbaoConnection: auth-bind handshake, ping heartbeats, pending correlator/send_and_wait, reconnection/backoff, shutdown semantics, and tests.
Message pipeline & channel facade src/openhuman/channels/providers/yuanbao/inbound.rs, src/openhuman/channels/providers/yuanbao/outbound.rs, src/openhuman/channels/providers/yuanbao/channel.rs	17-stage inbound middleware, OutboundSender APIs (text/image/file/upload via COS), YuanbaoChannel wiring including aliasing, heartbeat management, and message splitting.
Protocol codecs & utilities src/openhuman/channels/providers/yuanbao/wire.rs, proto.rs, proto_biz.rs, proto_constants.rs, types.rs, media.rs, cos.rs, ids.rs, splitter.rs, errors.rs, Cargo.toml	Add protobuf wire primitives (safe varints), ConnMsg/TIM codecs, biz-layer codecs, protocol constants, image parsing, COS HMAC-SHA1 signing/upload, ID shortening, markdown splitter, error enums, and add sha1 = "0.10".
Module organization & health check src/openhuman/channels/providers/mod.rs, src/openhuman/channels/mod.rs, src/openhuman/channels/commands.rs, docs/TEST-COVERAGE-MATRIX.md	Expose pub mod yuanbao, re-export YuanbaoChannel/YuanbaoConfig, integrate into doctor_channels health-checks, and add test-coverage matrix entry.

Estimated code review effort

🎯 5 (Critical) | ⏱️ ~120 minutes

Suggested labels

agent

Suggested reviewers

• graycyrus

🐰 A whisker-twitch of joy! The Yuanbao gold now flows
Through pipes of TypeScript, Rust, and proto-codes
Where tokens sign with HMAC-256 pride
And messages split fence-aware, far and wide
The garden grows—another channel blooms! 🌿✨

[coderabbitai]

Actionable comments posted: 6

🧹 Nitpick comments (5)

src/openhuman/channels/providers/yuanbao/config.rs (1)

143-161: 💤 Low value

Consider documenting the expected call order for apply_env_defaults and validate.

validate() requires ws_domain and api_domain (when no token) to be non-empty, but apply_env_defaults() fills these from env. If a caller validates before applying defaults, validation fails unexpectedly. The tests show the intended pattern (defaults applied first), but the API doesn't enforce or document it.

A defensive option would be to call apply_env_defaults internally at the start of validate, or add a doc comment clarifying the expected order.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/openhuman/channels/providers/yuanbao/config.rs` around lines 143 - 161,
The validate() method on the Yuanbao config currently fails if env-derived
fields (ws_domain, api_domain when token is absent) haven't been populated;
update validate to defensively call apply_env_defaults() at the start (or ensure
it returns/propagates any error) so env defaults are applied before checks, then
perform the existing checks and return YuanbaoError::Config as before;
alternatively, if you prefer not to mutate state during validation, add a clear
doc comment on validate mentioning callers must call apply_env_defaults() first
and reference the apply_env_defaults and validate functions in the docstring.

src/openhuman/channels/controllers/ops.rs (1)

180-194: 💤 Low value

Minor: Consider reusing an existing HTTP client.

verify_yuanbao_credentials creates a new reqwest::Client on each call. While acceptable for a one-time connect operation, the existing pattern elsewhere in the codebase typically shares a client for connection pooling benefits.

This is fine for now since credential verification is infrequent.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/openhuman/channels/controllers/ops.rs` around lines 180 - 194,
verify_yuanbao_credentials creates a fresh reqwest::Client for each call which
prevents reuse of connection pooling; change the function signature to accept a
shared reqwest::Client (or obtain the existing shared client) and pass that into
SignManager::new instead of constructing reqwest::Client::new() so
verify_yuanbao_credentials uses the shared client instance used elsewhere in the
codebase (refer to verify_yuanbao_credentials and SignManager::new(...) in the
diff).

src/openhuman/channels/providers/yuanbao/sign.rs (1)

261-261: 💤 Low value

Consider explicit handling when code field is missing.

At line 261, if the response JSON lacks a code field, it defaults to 0, which means the code proceeds as if successful. The subsequent check for data at line 263-270 provides a safety net, but an explicit check for a missing code field would make the intent clearer and fail faster with a more descriptive error.

This is a minor robustness improvement since the current flow still fails correctly.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/openhuman/channels/providers/yuanbao/sign.rs` at line 261, The current
extraction let code = json.get("code").and_then(|c| c.as_i64()).unwrap_or(0)
masks a missing "code" field by treating it as success; change the logic in the
handler (where variable code is created) to explicitly check whether
json.get("code") is Some and return a clear Err (or Err variant/early return)
when it's absent or not an integer, rather than falling back to 0, so callers of
this function (the response parsing around code and data) get a descriptive
failure instead of proceeding as if successful.

app/src/components/channels/__tests__/YuanbaoConfig.test.tsx (1)

82-95: ⚡ Quick win

Avoid className-coupled validation assertions.

This assertion depends on text-coral, so harmless style refactors can break the test. Assert the validation message count/content directly instead of CSS classes.

As per coding guidelines: “Prefer testing behavior over implementation details.”

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@app/src/components/channels/__tests__/YuanbaoConfig.test.tsx` around lines 82
- 95, The test currently filters elements by className ('text-coral') to detect
validation errors; update the assertions to avoid coupling to CSS by checking
the validation text/content directly: use screen.getAllByText(/AppID/) or
screen.queryAllByText(/AppID/) and assert on the returned array length (e.g.,
toBeGreaterThan(0) before input and toHaveLength(0) or toBe(0) after firing
fireEvent.change on the '元宝开放平台 AppID' field) and keep the expect that
channelConnectionsApi.connectChannel was not called; locate and modify the
assertions around screen.getAllByText, screen.queryAllByText, and the
fireEvent.change usage in YuanbaoConfig.test.tsx.

src/openhuman/channels/providers/yuanbao/splitter.rs (1)

6-7: 💤 Low value

Documentation mentions table-row handling but implementation doesn't track table state.

The module doc mentions "a Markdown table row (lines starting with |)" but safe_to_break only checks fence state. If table preservation was intentional, the state tracking is missing; otherwise, consider updating the doc comment.

Also applies to: 81-83

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/openhuman/channels/providers/yuanbao/splitter.rs` around lines 6 - 7, The
module doc claims splitter handles "a Markdown table row (lines starting with
`|`)" but the implementation (see function safe_to_break and surrounding state)
never tracks table state; either remove/update that sentence in the module
comment or implement table detection: add a boolean field (e.g., in the same
state struct used by safe_to_break) that is set when a line begins with '|' (and
cleared on blank lines or other block boundaries) and then update safe_to_break
to return false while in-table so table rows are preserved; pick one approach
and ensure you reference/modify the state struct and the safe_to_break function
accordingly.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@app/src/components/channels/YuanbaoConfig.tsx`:
- Around line 67-93: Add a synchronous re-entrancy guard to prevent rapid
repeated clicks from enqueueing duplicate actions: at the very start of
handleConnect check the local busy flag and return immediately if busy; do the
same for the other action handler (the disconnect/restart handler around lines
186–208). Keep the existing setBusy(true) usage but ensure the early busy check
is present before any async work or dispatch (and ensure setBusy(false) is
restored on completion/error so the guard is cleared).

In `@src/openhuman/channels/providers/yuanbao/channel.rs`:
- Around line 94-100: PipelineState is being created with config.bot_id at
startup so when config.bot_id is empty the echo-guard (SkipSelfMw) cannot match
self-sent messages after sign-token/auth-bind resolves the canonical bot id;
locate where sign-token/auth-bind yields the resolved bot id and update the
inbound pipeline with it by either calling a setter on the existing
PipelineState (e.g., PipelineState::set_from_account or set_bot_id) or by
replacing the pipeline with a new InboundPipeline(PipelineState::new(&config,
resolved_bot_id)); ensure the same PipelineState instance used by
InboundPipeline is updated (or pipeline swapped) after auth so SkipSelfMw can
match outbound echoes.
- Around line 294-296: After sending the shutdown signal via shutdown_tx, do not
immediately abort conn_task; instead await the connection task to allow
YuanbaoConnection::shutdown() to run and perform its cleanup so sender, pending,
and is_connected are updated; replace conn_task.abort() with awaiting the task's
JoinHandle (optionally with a timeout) and handle its Result, ensuring listen()
returns only after the task finishes.

In `@src/openhuman/channels/providers/yuanbao/connection.rs`:
- Around line 207-243: The reconnect attempt counter is never reset so
intermittent clean disconnects will exhaust max_attempts; after a successful
session (when self.connect_once(...) returns Ok(...) in the match) reset attempt
to 0 instead of only incrementing later. Locate the match on outcome in the
connection loop (connect_once, outcome, NO_RECONNECT_CLOSE_CODES) and set
attempt = 0 in the Ok arms (except where you return for no-reconnect close
codes) so cleanly closed/authenticated sessions restore the reconnect budget
before the backoff and subsequent sleep logic.

In `@src/openhuman/channels/providers/yuanbao/outbound.rs`:
- Around line 23-27: The send_* functions (e.g., send_group_message and
send_c2c_message) currently treat any frame with a matching msg_id as success;
after calling send_and_wait() they must inspect the business-layer response
payload (resp.data) and verify the inner response code/message before returning
success. Fix: after send_and_wait() in send_group_message and send_c2c_message,
decode or parse resp.data to extract the business fields (e.g., code and
message) using the appropriate decoder for that response (or parse JSON/proto
from resp.data), and if the inner code != 0 return/propagate an error containing
that code/message instead of success; apply the same check to the
heartbeat/send_*_heartbeat paths mentioned. Ensure you reference and use
send_and_wait, send_group_message, send_c2c_message (and the heartbeat send
functions) so the logic always validates resp.data.code before signaling
success.

In `@src/openhuman/channels/providers/yuanbao/wire.rs`:
- Around line 42-55: The varint decoder loop in wire.rs currently allows a
malformed 10th-byte varint to overflow/truncate; update the loop in the varint
decoding function (the block using variables i, pos, shift, byte, value and
returning YuanbaoError::ProtoDecode) to reject a 10th byte with any payload bits
other than the low bit: specifically, when shift == 63 (i.e. handling the 10th
byte) check if (byte & 0xFE) != 0 and return
Err(YuanbaoError::ProtoDecode("varint too long".into())) instead of shifting and
accepting those bits; keep the existing checks for truncated inputs and shift >=
64.

---

Nitpick comments:
In `@app/src/components/channels/__tests__/YuanbaoConfig.test.tsx`:
- Around line 82-95: The test currently filters elements by className
('text-coral') to detect validation errors; update the assertions to avoid
coupling to CSS by checking the validation text/content directly: use
screen.getAllByText(/AppID/) or screen.queryAllByText(/AppID/) and assert on the
returned array length (e.g., toBeGreaterThan(0) before input and toHaveLength(0)
or toBe(0) after firing fireEvent.change on the '元宝开放平台 AppID' field) and keep
the expect that channelConnectionsApi.connectChannel was not called; locate and
modify the assertions around screen.getAllByText, screen.queryAllByText, and the
fireEvent.change usage in YuanbaoConfig.test.tsx.

In `@src/openhuman/channels/controllers/ops.rs`:
- Around line 180-194: verify_yuanbao_credentials creates a fresh
reqwest::Client for each call which prevents reuse of connection pooling; change
the function signature to accept a shared reqwest::Client (or obtain the
existing shared client) and pass that into SignManager::new instead of
constructing reqwest::Client::new() so verify_yuanbao_credentials uses the
shared client instance used elsewhere in the codebase (refer to
verify_yuanbao_credentials and SignManager::new(...) in the diff).

In `@src/openhuman/channels/providers/yuanbao/config.rs`:
- Around line 143-161: The validate() method on the Yuanbao config currently
fails if env-derived fields (ws_domain, api_domain when token is absent) haven't
been populated; update validate to defensively call apply_env_defaults() at the
start (or ensure it returns/propagates any error) so env defaults are applied
before checks, then perform the existing checks and return YuanbaoError::Config
as before; alternatively, if you prefer not to mutate state during validation,
add a clear doc comment on validate mentioning callers must call
apply_env_defaults() first and reference the apply_env_defaults and validate
functions in the docstring.

In `@src/openhuman/channels/providers/yuanbao/sign.rs`:
- Line 261: The current extraction let code = json.get("code").and_then(|c|
c.as_i64()).unwrap_or(0) masks a missing "code" field by treating it as success;
change the logic in the handler (where variable code is created) to explicitly
check whether json.get("code") is Some and return a clear Err (or Err
variant/early return) when it's absent or not an integer, rather than falling
back to 0, so callers of this function (the response parsing around code and
data) get a descriptive failure instead of proceeding as if successful.

In `@src/openhuman/channels/providers/yuanbao/splitter.rs`:
- Around line 6-7: The module doc claims splitter handles "a Markdown table row
(lines starting with `|`)" but the implementation (see function safe_to_break
and surrounding state) never tracks table state; either remove/update that
sentence in the module comment or implement table detection: add a boolean field
(e.g., in the same state struct used by safe_to_break) that is set when a line
begins with '|' (and cleared on blank lines or other block boundaries) and then
update safe_to_break to return false while in-table so table rows are preserved;
pick one approach and ensure you reference/modify the state struct and the
safe_to_break function accordingly.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro

Run ID: 29cac06e-2dc3-404c-8717-90802e523b0f

📥 Commits

Reviewing files that changed from the base of the PR and between cc67177 and aa8fb8e.

⛔ Files ignored due to path filters (2)

• Cargo.lock is excluded by !**/*.lock
• app/src-tauri/Cargo.lock is excluded by !**/*.lock

📒 Files selected for processing (65)

• Cargo.toml
• app/src/components/channels/ChannelSelector.tsx
• app/src/components/channels/ChannelSetupModal.tsx
• app/src/components/channels/YuanbaoConfig.tsx
• app/src/components/channels/YuanbaoIcon.tsx
• app/src/components/channels/__tests__/ChannelSetupModal.test.tsx
• app/src/components/channels/__tests__/YuanbaoConfig.test.tsx
• app/src/components/channels/__tests__/YuanbaoIcon.test.tsx
• app/src/components/skills/skillIcons.tsx
• app/src/lib/i18n/chunks/ar-3.ts
• app/src/lib/i18n/chunks/ar-5.ts
• app/src/lib/i18n/chunks/bn-3.ts
• app/src/lib/i18n/chunks/bn-5.ts
• app/src/lib/i18n/chunks/de-3.ts
• app/src/lib/i18n/chunks/de-5.ts
• app/src/lib/i18n/chunks/en-3.ts
• app/src/lib/i18n/chunks/en-5.ts
• app/src/lib/i18n/chunks/es-3.ts
• app/src/lib/i18n/chunks/es-5.ts
• app/src/lib/i18n/chunks/fr-3.ts
• app/src/lib/i18n/chunks/fr-5.ts
• app/src/lib/i18n/chunks/hi-3.ts
• app/src/lib/i18n/chunks/hi-5.ts
• app/src/lib/i18n/chunks/id-3.ts
• app/src/lib/i18n/chunks/id-5.ts
• app/src/lib/i18n/chunks/it-3.ts
• app/src/lib/i18n/chunks/it-5.ts
• app/src/lib/i18n/chunks/ko-3.ts
• app/src/lib/i18n/chunks/ko-5.ts
• app/src/lib/i18n/chunks/pt-3.ts
• app/src/lib/i18n/chunks/pt-5.ts
• app/src/lib/i18n/chunks/ru-3.ts
• app/src/lib/i18n/chunks/ru-5.ts
• app/src/lib/i18n/chunks/zh-CN-3.ts
• app/src/lib/i18n/chunks/zh-CN-5.ts
• app/src/lib/i18n/en.ts
• app/src/store/__tests__/channelConnectionsSlice.test.ts
• app/src/store/channelConnectionsSlice.ts
• app/src/types/channels.ts
• docs/TEST-COVERAGE-MATRIX.md
• src/openhuman/channels/commands.rs
• src/openhuman/channels/controllers/definitions.rs
• src/openhuman/channels/controllers/ops.rs
• src/openhuman/channels/controllers/ops_tests.rs
• src/openhuman/channels/mod.rs
• src/openhuman/channels/providers/mod.rs
• src/openhuman/channels/providers/yuanbao/channel.rs
• src/openhuman/channels/providers/yuanbao/config.rs
• src/openhuman/channels/providers/yuanbao/connection.rs
• src/openhuman/channels/providers/yuanbao/cos.rs
• src/openhuman/channels/providers/yuanbao/errors.rs
• src/openhuman/channels/providers/yuanbao/ids.rs
• src/openhuman/channels/providers/yuanbao/inbound.rs
• src/openhuman/channels/providers/yuanbao/media.rs
• src/openhuman/channels/providers/yuanbao/mod.rs
• src/openhuman/channels/providers/yuanbao/outbound.rs
• src/openhuman/channels/providers/yuanbao/proto.rs
• src/openhuman/channels/providers/yuanbao/proto_biz.rs
• src/openhuman/channels/providers/yuanbao/proto_constants.rs
• src/openhuman/channels/providers/yuanbao/sign.rs
• src/openhuman/channels/providers/yuanbao/splitter.rs
• src/openhuman/channels/providers/yuanbao/types.rs
• src/openhuman/channels/providers/yuanbao/wire.rs
• src/openhuman/channels/runtime/startup.rs
• src/openhuman/config/schema/channels.rs