Recover skill runtime policy/admin services and Skill Router v2 helpers so
admins can toggle manifest routing and inspect the platform skill catalog.
Co-authored-by: Cursor <cursoragent@cursor.com>
Page Data questionnaires were returning publication slugs to users; password-mode
pages always fell back to /u/ routes. Prefer MindSpace paths at publish time and
rewrite mistaken delivery links in WeChat and Portal chat.
Co-authored-by: Cursor <cursoragent@cursor.com>
When schedule_create_item includes a start time and reminder intent, create the
matching h5_schedule_reminders row in the same tool call so 10:00 pushes are not
lost. Also reconcile Goose SSE request ids with the portal agent-run gate to keep
MindSpace chat streaming reliable.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add chat file/image upload UX, attachment proxying, vision thumbnails, and per-turn image scoping so agents only use the current upload. Extend MindSpace asset context, billing token state, OA/scenario verify scripts, and related runtime config.
Co-authored-by: Cursor <cursoragent@cursor.com>
Keep online publication snapshots aligned when public/*.html changes, default static page binds to public access with MindSpace delivery URLs, and split browser-safe workspace helpers out of the Vite import chain.
Co-authored-by: Cursor <cursoragent@cursor.com>
Pre-bundle Chart.js, auto-rewrite common CDN references at publish and serve time, and block unknown external script src during publication scans so interactive dashboards keep working under CSP.
Co-authored-by: Cursor <cursoragent@cursor.com>
Wire chat intent routing with direct_chat on regular sessions, skill-selected
short-circuit to Agent, memory light/heavy intervention tiers, and fix direct
chat UI stuck streaming after completion.
Co-authored-by: Cursor <cursoragent@cursor.com>
Inject serve-time CSS so author styles cannot hide TKMind · 智趣 with
low opacity, which made pages like weather.html look brandless.
Co-authored-by: Cursor <cursoragent@cursor.com>
Expose generate_docx in mindspace-sandbox-mcp so agents can write public/*.docx
before linking HTML download pages, with tests mirroring the Mark summary flow.
Co-authored-by: Cursor <cursoragent@cursor.com>
Diagnosed via production session logs (RDS + shared goose PG) that some
users' page-generation attempts drifted: the agent used apps__create_app
(no public URL) then fabricated a placeholder domain when asked for a
link, and separately burned many turns scraping unreachable/anti-bot
search engines. Steer both flows toward the already-working path instead
of banning tools outright:
- apps__create_app is fine for designing/previewing a page, but the
agent must still write_file the result into public/*.html per the
static-page-publish skill and return a link built from the real
public URL template (no invented domains)
- for real-world lookups, load_skill('web') first and prefer
Bing/360 over retrying google.com or hammering anti-bot sites
Also folds in generate_docx guidance for Word-download attachments
(sandbox-fs tool) that landed in the same files during this pass.
Co-Authored-By: Claude <noreply@anthropic.com>
Add a MindSpace public HTML checker for missing companion docx/PDF paths, wire it into runtime packaging and release-portal-runtime-prod.sh, and document the workflow for agents and ops.
Co-authored-by: Cursor <cursoragent@cursor.com>
Replace fixed ackText with a rule-based AckProvider that picks
response templates by message type and intent (translate, summary,
rewrite, poster, ppt, mindmap, code, search, schedule). Pure sync,
zero I/O, auto-falls back to config.ackText on any error.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Split platform admin and ops APIs into standalone admin-server.mjs with network guards; simplify billing to RMB token pricing, refactor user auth, and add rsync deploy plus local-test scripts and docs.
Co-authored-by: Cursor <cursoragent@cursor.com>
Introduce AI image generation with chat shortcut and agent API, improve MindSpace chat-to-page save resolution, and seed Plaza covers with production deploy scripts.
Co-authored-by: Cursor <cursoragent@cursor.com>
Introduce page edit sessions with draft preview and patch API, chat skill picker, user memory profile, h5ApiBase resolution, voice WAV transport, and scripts for 105/g2 deployment and Plaza local dev.
Co-authored-by: Cursor <cursoragent@cursor.com>
Track application source and tests; exclude local env, user workspaces, and runtime data via .gitignore.
Co-authored-by: Cursor <cursoragent@cursor.com>