docs(i18n): sync API server discovery surfaces

[r266-tech]

Summary

• preserve the existing, more detailed English Sessions API and skills/toolsets sections instead of duplicating them
• add localized equivalents of those current sections to the zh-Hans API-server guide
• update both /v1/capabilities JSON examples with the live skills_api and four session_* feature flags

The localized Sessions API table covers all nine registered routes, plus fork/chat examples, capability discovery, inline-image support, and the read-only skills/toolsets endpoints.

Validation

• verified all five feature flags are true in gateway/platforms/api_server.py
• verified the nine /api/sessions routes and both /v1/skills / /v1/toolsets registrations
• parsed both updated capability snippets as JSON
• confirmed no overlapping open localization PR or target-file relocation
• git diff --check

Docs-only change; server behavior is unchanged. This is the targeted salvage requested in the maintainer review.

[teknium1]

Thanks for documenting these externally consumable API surfaces. Current main already has a more detailed English implementation of the endpoint documentation, so the English portion needs a targeted salvage rather than a direct application.

Problems

• 8b6beaab5 added the current English Sessions API and discovery sections at website/docs/user-guide/features/api-server.md:332-377; the PR's English endpoint additions overlap that material.
• The current /v1/capabilities example at website/docs/user-guide/features/api-server.md:205-219 still omits live skills_api and session_* flags exposed by gateway/platforms/api_server.py:1526-1545.

Suggested changes

• Preserve the zh-Hans coverage: its API-server page still jumps from Jobs to System Prompt Handling at website/i18n/zh-Hans/docusaurus-plugin-content-docs/current/user-guide/features/api-server.md:316.
• Update both capability examples from the current server contract, while avoiding duplicate English endpoint sections.

Automated hermes-sweeper review.

[r266-tech]

Reworked per the salvage guidance:

• removed the now-duplicate English endpoint additions and retained current main’s richer English sections
• translated those current Sessions API and skills/toolsets discovery sections into zh-Hans
• updated both capability examples with skills_api, session_resources, session_chat, session_chat_streaming, and session_fork

I verified the flags and all registered routes from gateway/platforms/api_server.py; both JSON snippets parse and git diff --check passes.

[GottZ]

This was generated by AI during triage.

Summary

Two PRs touch the API-server documentation, but they address different causes: #33283 synchronizes capability discovery and missing zh-Hans session/skills documentation, while #62216 attempted comma-separated API-server master keys across authentication code, configuration, and docs.

Related pull requests

• #33283 related — (+59/-2) — merge: The revised diff follows the contributor’s keep_open salvage guidance by preserving the richer English endpoint sections, adding the five live capability flags to both examples, and filling the missing zh-Hans Sessions API and skills/toolsets coverage.
• #62216 [closed] related — (+58/-13) — keep closed: Despite the earlier keep_open review on #62216, the diff drops blank key segments, leaves tests referencing the removed _api_key, and would regress newer encoded-bytes and fail-closed authentication work; the later contributor decision also documents that multiplicity belongs in scoped credential-pool/token designs tracked by #44615 and #64697, not comma-separated equivalent master keys.

Suggested consolidation

Merge #33283 as the targeted, review-aligned documentation synchronization. Keep #62216 closed rather than treating it as a duplicate: it addresses a separate authentication design, has concrete correctness and regression issues, and conflicts with the documented repository direction toward scoped credentials.

Cross-PR triage: Reviewed 2 pull requests and 0 issues in this complex. Each diff was read against this issue; Assessment working set: 12 kB of PR diffs, 2 kB of issue/PR text, 4 kB of discussion (6 comments), 0 verify verdicts. verdicts reflect diff content, not PR titles. Part of an automated triage batch.