docs: chats improvement plan (multiplayer, turn model, memory, documents, insights)#757
Open
spashii wants to merge 2 commits into
Open
docs: chats improvement plan (multiplayer, turn model, memory, documents, insights)#757spashii wants to merge 2 commits into
spashii wants to merge 2 commits into
Conversation
…ory, documents, insights, support escalation Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
spashii
added a commit
that referenced
this pull request
Jul 2, 2026
Phase 0 of `echo/docs/chats-improvement-plan.md` (PR #757): the customer-visible brand fixes. - Rewritten SYSTEM_PROMPT (no "AI", honest "I don't know", scope/turn-instruction sections ready for the scoping+templates work, never-claims-applied). Citation format byte-identical (frontend regex). - Title prompt names the subject, never the tool. - Worker: no tool-name leak into chat, single honest tool-limit message, no English "Next steps:" label. Agent tests 43/43, worker tests 19/19 (assertions updated to the new intended copy). 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
spashii
added a commit
that referenced
this pull request
Jul 2, 2026
Frontend half of Phase 0/1 of the chats plan (#757), aimed at non-technical hosts. - **One simple surface:** with agentic enabled, new and mode-less chats start automatically — no mode selector, no jargon. Selector kept for flag-off environments (prod until launch). - **Suggested-changes card**: friendly portal-editor labels, per-field old → new with honest empty states, **editable proposed values** (fine-tune before applying), switches for booleans, and **stateless applied detection** — the card compares live project values so reloads show 'applied' truthfully (fixes the 'did it apply?' confusion). - Removes 'Run status: completed', fixes the scroll-button overlap + off-brand variant, chip mid-word clipping, placeholder/empty-state copy, lowercase dembrane, i18n-wraps tool headlines, 12px type floor. Verified: tsc, biome (repo-level, error-clean), lingui extract+compile, vite build. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
spashii
added a commit
that referenced
this pull request
Jul 2, 2026
Host feedback: the per-tool 'Reading a skill / Searching the documentation — DONE' rows look like a debug log, and 'reading a skill' isn't intuitive. Consecutive tool activities now fold into one subtle strip — live step while working, 'Worked through N steps' when done, expandable to the individual steps with raw input/output still available behind a per-step details toggle (kept for devs, out of the host's way). Messages and suggestion cards break a group so the thread reads as a conversation. First slice of the chat redesign (plan #757). Verified: tsc, biome (repo-level), lingui, vite build. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
spashii
added a commit
that referenced
this pull request
Jul 2, 2026
Host feedback: 'as soon as I hit enter, my message does not immediately show up is a bit worrying.' The message now appears instantly (optimistic echo), and is replaced by the persisted one once it streams back (deduped by content), or cleared on submit error. Second slice of the chat redesign (#757). Verified: tsc, biome (repo-level), vite build. 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The consolidated plan from today's production pass and product direction: agentic-only mode, conversation scoping, editable/immutable suggestions, multiplayer (with concerns), ack-steered durable turns (LangGraph checkpointing now, Temporal later, no Bun), action registry to UI parity, omnipresent chat, three-level memory (workspace/project/user), document uploads (extract/summarize/grep), anonymized usage insights, and reachOutToDembrane() support escalation into Slack for Sam.
Ends with 8 decisions to make. Grounded in live echo-next data and the three production-readiness assessments.
🤖 Generated with Claude Code