main
14240 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Maxim
|
25817605cc |
fix(reskinnable-demo): put real data behind the banking report, and fix the six defects it was hiding (#6378)
Fixes the six defects a 12-agent review found in the banking skin, by
fixing the thing that caused most of them: the skin had **three
disagreeing answers to "what did we spend"**.
| Source | Total | Drove |
|---|---|---|
| `charges-data.ts` fixture (45 rows, Apr–Jun) | $632,806 | the Charges
page |
| `seed.json` ledger (4 rows, **Apr–May only**) | $30,089 | the report's
**charts** |
| static `policies[].spent` | $137,000 | the report's **KPI** |
Every one of those numbers appeared in the product, and they were never
reconciled.
## The headline defect
`SpendingTrendChart` substituted a hard-coded `[3200, 4100, 3600, 5200,
4800, 6400]` / Jan–Jun series whenever fewer than three distinct months
were present. Intended as an empty state — but the ledger spanned
exactly **two** months, so the fallback was the **default path**. The
report's "Spend over time" showed six invented figures, roughly 20×
smaller than the total printed directly above them, under a card whose
own docstring reads:
> Every number is computed here from the live ledger … so a report can
never quote a figure the app disagrees with.
Reproduced in the running app before fixing (`POST
/api/banking/v1/reports`, no agent needed), and there's a nasty
interaction worth knowing: attaching an invoice dates a synthetic
transaction *today*, supplying a third month, so **attaching an invoice
masked the bug**. A test written casually would sit in the masked state
and pass.
## The fix: one ledger
The 45 charges now live in `seed.json` as real transactions across
**Apr/May/Jun**, and the Charges page reads them over REST like every
other surface.
**Team and policy are now different axes.** A charge belongs to one of
seven org teams; a policy is one of three budget envelopes (Technology /
Go-to-Market / G&A) and several teams share one. They were a single
`ExpenseRole` enum — which is exactly why covering every team meant
choosing between a seven-slice donut and discarding real charges.
`ExpenseRole` still types a member's own team; `PolicyType` types the
envelopes, joined by `policyForTeam`.
**`policies[].spent` is derived from approved charges on every read.**
It could no longer disagree with the charts — and it also now *moves*:
approving a charge previously left `spent` untouched, so the budget
never reflected the approval and the over-limit gate kept comparing
against a stale figure. Verified live: approving a \$960 charge moved
`spent` by exactly \$960.
**Over-limit is derived-only.** A charge no longer *stores*
`over-limit`; the Charges badge resolves through `withOverLimit`, the
same rule the report uses.
## The six review findings
| | Defect | Fix |
|---|---|---|
| a1 | report charted fabricated spend | charts real months; explicit
empty state at zero |
| a2 | donut its own docs said would collapse | real \$533k base — the
\$900k invoice that hit **89%** now reaches **73%**; 89% would need
~\$2.8M |
| a3 | `report` matched inside "quarterly report" | anchored both sides,
whole words |
| a4 | `as Transaction` laundered a nullable `policyId` | cast removed;
compiler checks it |
| a5 | comment claimed additions "have no policyId" | corrected — three
lines from the code that sets it |
| a6 | `?sort=banana` lit the "active" tint | params validated; unknown
reads as unset |
All six existed **identically in `examples/showcases/banking/`** — they
came from the upstream PRs this skin replayed, not from the port. Scoped
to the skin per review; banking still carries them.
## The scripted demo is unchanged by construction
- the four demo-load-bearing transactions survive byte-identical
- over-limit is still **exactly 3 charges / \$30,000**
- AWS \$15,000 still derives over-limit for the teach-mode pill
- Delta Airlines is still the only Delta charge (the fixture's
near-duplicate "Delta Air Lines" became United Airlines)
- all four status badges still appear (`Amazon Business` is kept
pending, under its policy's headroom, so a plain **Pending** chip
survives)
The donut goes 44/40/16 → **48/27/24** and is relabelled "Spend by
policy", which is what it reads.
## Verification
```
nx build react-core,a2ui-renderer,core,runtime,shared exit 0
tsc --noEmit 0 errors
vitest 239 passed (was 215)
eslint 0 problems
```
Plus live checks against a running server: 49 transactions over 3
months, derived `spent` tracking an approval, over-limit holding at
3/\$30,000.
The 24 new tests are confirmed **red** against the old code —
`parseSort("banana")` returned `"banana"`, `parseTop("-5")` returned
`-5`, the seed spanned two months, no row carried a team, and `spent`
was stored.
## Not in this PR
A "tool replay-guard sweep" (4 items, `navigateToPageAndPerform` and
three approval tools missing the resolved-state guard `showCharges` has)
and ~13 subject-neutral items are captured as follow-ups.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
|
||
|
Maxim
|
8829633e29 | Merge branch 'main' into fix/banking-skin-real-data | ||
|
Maxim
|
addbe2888d |
test(reskinnable-demo): lock in the single-ledger invariants
Two new suites, each written so it would have failed against the previous code
rather than merely describing the new behaviour:
- derived-spend: policies[].spent equals the sum of approved charges, excludes
pending/flagged (so the over-limit check cannot double-count the charge it is
gating), and agrees between policies() and findPolicy(). Plus the properties
the demo depends on — at least three distinct months (the condition whose
absence made the trend chart fabricate), AWS $15,000 still deriving
over-limit for the teach-mode pill, exactly one Delta charge, and a team and
category on every ledger row.
- charges-data: parseSort and parseTop reject the values that previously slipped
through as valid, and toChargeRow's over-limit projection only applies to
pending charges.
Confirmed red against the old code: parseSort("banana") returned "banana",
parseTop("-5") returned -5, the seed spanned two months, no row carried a team,
and spent was a stored field.
The four existing fixtures move from ExpenseRole to PolicyType for policy
`type`, following the team/policy split.
|
||
|
Maxim
|
d7e7c8eaf4 |
fix(reskinnable-demo): stop the alert marker firing on ordinary notes
The regex that decides whether a transaction note gets a red-alert prefix was anchored on the left only, so `report` matched inside "reporter" and "quarterly report" and `disput` inside "disputation". A note reading "attached to the quarterly report" was served a fraud marker. Anchored on both sides and switched from stems to whole words, verified against the four phrasings that previously false-positived. |
||
|
Maxim
|
05390dc625 |
fix(reskinnable-demo): stop the banking report showing invented figures
Four defects in the report's charts, all reported by review and all reproduced in the running app before fixing. - SpendingTrendChart substituted a hard-coded [3200, 4100, 3600, 5200, 4800, 6400] Jan-Jun series whenever fewer than three months were present. Intended as an empty state, it was the DEFAULT path: the seeded ledger spanned two months, so the report's "Spend over time" always showed six invented numbers — roughly 20x smaller than the total printed directly above them — under a card whose own contract says every number comes from the live ledger. It now charts whatever months exist, with a real empty state at zero. - SpendBreakdownChart's docstring said the report must use SpendByTeamBars instead, "because an attached invoice can push one team to ~96% and a donut cannot survive that", while the report rendered the donut anyway. The warning was real but its cause was the thin ledger, not the chart: against the old $137,000 base a $900,000 invoice took one slice to 89%. Against the real ~$533,000 base the same invoice reaches 73%, and 89% would need ~$2.8M. Robust because the data is real, not because a floor was added to the arc. - augmentForReport built its synthetic transactions behind an `as Transaction` cast that was hiding a real hole: `policyId` came from an `?.id` lookup, so it could be undefined where the field is a required string. The cast is gone and the compiler checks it. Additions also now resolve their model-authored team to a policy envelope through `policyForTeam` rather than comparing a team name to a policy name, with one "Unattributed" segment for unmappable names. - A comment inside TopChargesChart claimed document-sourced charges "have no policyId" — three lines above the code that gives them one. The donut column is relabelled "Spend by policy", which is what it reads. |
||
|
Maxim
|
274d46ffcb |
feat(reskinnable-demo): put one real ledger behind the banking skin
The skin held three disagreeing answers to "what did we spend": a 45-row Charges fixture ($632,806), a 4-row seeded ledger ($30,089 across two months), and static policy totals ($137,000). Each surface read a different one, so they drifted silently — and because the ledger spanned only two months, the report's trend chart fell back to a hard-coded series and showed invented figures under a card that promises live numbers. Now there is one ledger. The 45 charges live in seed.json as real transactions across Apr/May/Jun, and the Charges page reads them over REST like every other surface. - Splits team from policy. A charge belongs to one of seven org teams; a policy is one of three budget envelopes (Technology / Go-to-Market / G&A) and several teams share one. These were a single `ExpenseRole` enum, which is why the two axes read as one thing and why covering every team meant either a seven-slice donut or discarding real charges. `ExpenseRole` still types a member's own team; `PolicyType` types the envelopes, joined by `policyForTeam`. - Derives `policies[].spent` from approved charges on every read, so it can no longer disagree with the charts. It also now MOVES: approving a charge previously left `spent` untouched, so the budget never reflected the approval and the over-limit gate kept comparing against a stale figure. - Makes over-limit derived-only. A charge no longer stores "over-limit"; the Charges table resolves the badge through `withOverLimit`, the same rule the report uses, so the two cannot disagree. - Validates the `?sort=` and `?top=` params. `?sort=banana` used to be cast straight to a SortKey and lit the control's "active" tint while the table silently sorted by the default; `?top=-5` reached `slice(0, -5)` and dropped the LAST five rows, inverting top-N. The scripted demo is unchanged by construction: the four demo-load-bearing transactions survive byte-identical, over-limit is still exactly three charges totalling $30,000, AWS $15,000 still derives over-limit for the teach-mode pill, and Delta Airlines is still the only Delta charge (the fixture's near -duplicate "Delta Air Lines" became United Airlines). |
||
|
Tyler Slaton
|
c236e33978 |
chore: release monorepo v1.66.2 (#6376)
🚀 pkg-pr-new / build (push) Has been cancelled
## Release monorepo v1.66.2 **Scope:** `monorepo` | **Bump:** `patch` --- ### How this release process works 1. **This PR was created automatically** by the "release / create-pr" workflow. It bumped the `monorepo` packages to `1.66.2` and generated AI-enhanced release notes. 2. **CI runs on this PR** — the full test suite (unit tests, lint, type checks, build) must pass before merging. This is the review gate. 3. **Review the release notes** in `release-notes.md` in this PR. If a Notion draft was created, you can edit the release notes there before merging. 4. **When this PR is merged**, the `release / publish` workflow automatically: - Builds all packages - Publishes the `monorepo` packages to npm at version `1.66.2` - Creates git tag `monorepo/v1.66.2` - Creates a GitHub Release with the final release notes ### Before merging - [ ] CI is green (tests, lint, types, build) - [ ] Version bumps look correct - [ ] Release notes are accurate (edit in Notion if a draft was created) --- > **Do not merge until CI is fully green.** The full test suite runs automatically on this PR.v1.66.2 |
||
|
tylerslaton
|
53b772552f | chore: release monorepo v1.66.2 | ||
|
Tyler Slaton
|
cc9b74bdb1 |
chore: release channels v0.7.3 (#6375)
🚀 pkg-pr-new / build (push) Has been cancelled
## Release channels v0.7.3 **Scope:** `channels` | **Bump:** `patch` --- ### How this release process works 1. **This PR was created automatically** by the "release / create-pr" workflow. It bumped the `channels` packages to `0.7.3` and generated AI-enhanced release notes. 2. **CI runs on this PR** — the full test suite (unit tests, lint, type checks, build) must pass before merging. This is the review gate. 3. **Review the release notes** in `release-notes.md` in this PR. If a Notion draft was created, you can edit the release notes there before merging. 4. **When this PR is merged**, the `release / publish` workflow automatically: - Builds all packages - Publishes the `channels` packages to npm at version `0.7.3` - Creates git tag `channels/v0.7.3` - Creates a GitHub Release with the final release notes ### Before merging - [ ] CI is green (tests, lint, types, build) - [ ] Version bumps look correct - [ ] Release notes are accurate (edit in Notion if a draft was created) --- > **Do not merge until CI is fully green.** The full test suite runs automatically on this PR.channels/v0.7.3 |
||
|
tylerslaton
|
b95a43e254 | chore: release channels v0.7.3 | ||
|
Mark
|
c1adb14c48 |
chore(showcase): pin opentelemetry-api<1.44 for pydantic-ai v1 (#6374)
`showcase/integrations/pydantic-ai` cannot be installed from scratch today. A clean `pip install -r requirements.txt` produces an agent that dies on import: ``` ModuleNotFoundError: No module named 'opentelemetry._events' ``` pydantic-ai 1.0.18 imports `opentelemetry._events`. That module was removed in **opentelemetry-api 1.44.0** (last present in 1.43.0). pydantic-ai declares an unbounded `opentelemetry-api>=1.28.0`, so a fresh resolve picks 1.44.0 and the import fails. ## Why nothing was red `Dockerfile:22-24` copies **only** `requirements.txt` before `pip install`, so that layer's cache key depends on nothing else. The agent image has been reusing a pip layer baked before 1.44.0 shipped (2026-07-16) — including a successful build earlier today. The failure was masked by Docker layer caching, not absent. Any change to `requirements.txt`, a cache eviction, or a `--no-cache` build surfaces it. Note that this PR busts that cache by definition, so CI's image build is a genuine fresh-resolve test of the fix rather than a cached pass. ## Verification Run with `pip` in a clean 3.12 venv, matching how the Dockerfile installs: ``` RED (main) opentelemetry-api 1.44.0 -> import pydantic_ai raises ModuleNotFoundError GREEN (this branch) opentelemetry-api 1.43.0 -> import pydantic_ai + pydantic_ai.ag_ui OK ``` Only the ceiling is added; `opentelemetry-sdk` follows to 1.43.0 on its own, so a second pin isn't needed. No resolution conflict with `logfire>=4.10.0`. Docker and the `--d6` probe stack aren't available in my environment, so the mandatory value-test per `showcase/AGENTS.md` rule 4 has not been run locally — CI's image build and the dojo cells are the real gate here. ## Scope One line plus a comment recording why it exists and when to remove it. The comment matters: an undocumented pin is what caused the sibling `starlette==0.45.3` rot in #6363, where a correct-when-written pin outlived its reason and silently capped pydantic-ai a full major. The ceiling should come off when this package moves to pydantic-ai v2, which requires `opentelemetry-api>=1.28.0` without needing `_events`. |
||
|
Tyler Slaton
|
224dfb0d8c |
fix(channels): reconnect after clean gateway close (#6371)
## Summary - replace Phoenix's retained closed transport when a live managed session receives an unexpected WebSocket close with code 1000 - reconnect and rejoin through the existing Phoenix channel lifecycle - preserve intentional session disconnects as terminal - cover the clean-close recovery path with a regression test ## Root cause Phoenix 1.8.4 does not schedule its reconnect timer for close code 1000. CopilotKit still transitioned the managed session to `reconnecting`, so the runtime reported that Phoenix was retrying indefinitely even though Phoenix never created another transport. ## Impact A brief gateway interruption that cleanly closes an established socket can now recover once the gateway is healthy, instead of leaving the runtime stuck and repeatedly logging the managed-session-down warning. ## Verification - `pnpm nx run @copilotkit/channels-intelligence:test` — 188 tests passed - `pnpm nx run @copilotkit/channels-intelligence:check-types` — passed - `pnpm nx format:check --files=packages/channels-intelligence/src/realtime-gateway.ts,packages/channels-intelligence/src/realtime-gateway.test.ts` — passed - repository pre-commit test, package validation, and lint gates — passed |
||
|
Tyler Slaton
|
7e30957976 | fix(channels): reconnect after clean gateway close | ||
|
Mark
|
80fa407978 |
chore(showcase): pin opentelemetry-api<1.44 for pydantic-ai v1
pydantic-ai 1.0.18 imports opentelemetry._events, removed in opentelemetry-api 1.44.0. Its own floor is an unbounded opentelemetry-api>=1.28.0, so a fresh resolve of this package's requirements picks 1.44.0 and import pydantic_ai fails with ModuleNotFoundError. The agent image kept building because the Dockerfile copies only requirements.txt before pip install, so that layer stayed cached from before 1.44.0 shipped (2026-07-16). Any requirements change or cache eviction would have surfaced it. Verified with pip in a clean 3.12 venv, as the Dockerfile installs: before: opentelemetry-api 1.44.0, import pydantic_ai raises after: opentelemetry-api 1.43.0, import pydantic_ai OK |
||
|
Mike Ryan
|
3eaeeb19ce |
fix(channels): preserve Slack provider diagnostics (#6373)
## Summary - serialize Slack Carousel cards through elements and validate Card and Carousel payloads before provider calls - keep bounded provider diagnostics from ChannelDeliveryError through RUN_ERROR and ChannelCanonicalRunError - log low-cardinality error fields without validation messages or provider bodies - move best-effort Slack status cleanup failures to debug logs ## Why Invalid nested Slack Card payloads failed at the provider with no useful path at the application boundary. This change catches known Card and Carousel shape errors locally and preserves safe JSON pointers when Slack rejects a payload. ## Compatibility The new error details are optional. Existing consumers keep their current behavior, so no protocol version bump is needed. ## Validation - NX_DAEMON=false pnpm nx run-many -t test,check-types,build --projects=@copilotkit/channels-core,@copilotkit/channels-intelligence,@copilotkit/channels-slack,@copilotkit/runtime --skip-nx-cache - pnpm lint - pnpm check:channel-native-catalogs - targeted oxfmt and oxlint checks - git diff --check - pre-commit test, publint, and attw checks for affected packages Companion Intelligence PR: https://github.com/CopilotKit/Intelligence/pull/759 |
||
|
Mark
|
cdd605c672 |
docs(pydantic-ai): port integration docs and demos to Pydantic AI v2 (#6367)
This pull request was posted by Claude Code using claude-opus-5 on behalf of David. David has not reviewed this diff line by line. Closes https://github.com/CopilotKit/CopilotKit/issues/6363 `Agent.to_ag_ui()`, `AGUIApp` and the whole `pydantic_ai.ag_ui` module were removed in Pydantic AI v2. The docs installed pydantic-ai unpinned, so following the quickstart today gets 2.22.0 and fails twice: first at resolution (`starlette==0.45.3` conflicts with the `>=0.46.2` the `ag-ui` extra requires), then at `AttributeError`. ## What changed **8 doc pages** under `showcase/shell-docs/src/content/docs/integrations/pydantic-ai/` (`quickstart.mdx`, `quickstart/pydantic-ai.mdx`, `human-in-the-loop.mdx`, `human-in-the-loop/agent.mdx`, `generative-ui/tool-rendering.mdx`, and the three `shared-state/` pages): - the agent is served from a Starlette route via `AGUIAdapter.dispatch_request(request, agent=agent)` - `StateDeps` imports move from `pydantic_ai.ag_ui` to `pydantic_ai.ui` - install commands exact-pin `pydantic-ai-slim[ag-ui,openai]==2.22.0` and `ag-ui-protocol==0.1.19`, matching the starter fleet, plus `starlette>=0.46.2` since the snippets import Starlette directly **Per-request deps.** Every stateful snippet builds `StateDeps` inside the request handler: ```python async def run_agent(request: Request) -> Response: return await AGUIAdapter.dispatch_request( request, agent=agent, deps=StateDeps(AgentState()) ) ``` `dispatch_request` validates the client's state into `deps.state` (`pydantic_ai/ui/_adapter.py`, `run_stream_native`), so a module-level instance shared across requests lets concurrent runs clobber each other. The old `to_ag_ui(deps=...)` snippets all did this. **`examples/canvas/pydantic-ai`** — `requirements.txt` pinned, `agent/agent.py` ported, README corrected. **`examples/showcases/pydantic-ai-todos`** — `pyproject.toml` pinned and `uv.lock` regenerated (it was still resolving 1.0.10), `agent/main.py` ported, `src/agent.py` and `src/tools.py` imports moved, README and `src/app/api/copilotkit/route.ts` comments corrected. **`skills/copilotkit-integrations`** — beyond the issue's file list: `SKILL.md`, `sources.md` and `references/integrations/pydantic-ai.md` also taught `to_ag_ui()`. Same rot, same fix. ## Verified by execution The reason these docs rotted is that nothing runs them, so everything below was actually run, not read. - Both install commands were run verbatim in throwaway environments. `uv add 'pydantic-ai-slim[ag-ui,openai]==2.22.0' 'ag-ui-protocol==0.1.19' 'starlette>=0.46.2' uvicorn` and the `pip install` equivalent both resolve, landing pydantic-ai-slim 2.22.0, ag-ui-protocol 0.1.19, starlette 1.3.1. - Every ```python fence on the 8 doc pages was extracted, `exec`'d, and driven with a real `RunAgentInput` POST through `starlette.testclient.TestClient` with the model overridden to `TestModel`. All 8 return 200 `text/event-stream` with a `RUN_STARTED` ... `RUN_FINISHED` sequence and no `RUN_ERROR`. - The canvas agent was installed from its `requirements.txt` and driven the same way: 200, SSE, `RUN_STARTED` ... `TOOL_CALL_*` ... `STATE_SNAPSHOT` ... `RUN_FINISHED`. - The todos agent was installed with `uv sync --frozen` from the regenerated lock and driven the same way. Two sequential requests, one seeding a todo and one sending empty state, each saw only their own state, confirming the per-request deps actually isolate. Not executed: the Next.js frontends and the docs site build (no `node_modules` in this checkout). The TypeScript edits are comment-only. ## Deliberately out of scope `showcase/integrations/pydantic-ai` is left on its v1 fleet pin. It is 418 files, 19 mounts and 190 e2e specs, and CopilotKit said they will take it as https://github.com/CopilotKit/CopilotKit/issues/6364. The dojo and the docs therefore diverge until that lands. The CI guard from the issue's last acceptance criterion is not built here. A proposal for it is posted on https://github.com/CopilotKit/CopilotKit/issues/6363 for the team to own. Two pre-existing malformed code fences were fixed in passing, because leaving them meant the ported snippets still would not run: `quickstart/pydantic-ai.mdx` and `shared-state/predictive-state-updates.mdx` each had TypeScript embedded inside an unterminated ```python fence. The TypeScript now sits in its own fence. Overlaps with https://github.com/CopilotKit/CopilotKit/pull/6355, which ports `examples/integrations/pydantic-ai`. No file overlap. |
||
|
Mike Ryan
|
80370b5ecb | fix(channels): preserve provider diagnostics | ||
|
Mike Ryan
|
ef531035e6 | fix(channels-slack): validate card payloads locally | ||
|
Mike Ryan
|
79080816cd | fix(channels-slack): serialize carousel cards as elements | ||
|
Tyler Slaton
|
aae4c0bd5d |
chore: release channels v0.7.2 (#6372)
🚀 pkg-pr-new / build (push) Has been cancelled
## Release channels v0.7.2 **Scope:** `channels` | **Bump:** `patch` --- ### How this release process works 1. **This PR was created automatically** by the "release / create-pr" workflow. It bumped the `channels` packages to `0.7.2` and generated AI-enhanced release notes. 2. **CI runs on this PR** — the full test suite (unit tests, lint, type checks, build) must pass before merging. This is the review gate. 3. **Review the release notes** in `release-notes.md` in this PR. If a Notion draft was created, you can edit the release notes there before merging. 4. **When this PR is merged**, the `release / publish` workflow automatically: - Builds all packages - Publishes the `channels` packages to npm at version `0.7.2` - Creates git tag `channels/v0.7.2` - Creates a GitHub Release with the final release notes ### Before merging - [ ] CI is green (tests, lint, types, build) - [ ] Version bumps look correct - [ ] Release notes are accurate (edit in Notion if a draft was created) --- > **Do not merge until CI is fully green.** The full test suite runs automatically on this PR.channels/v0.7.2 |
||
|
tylerslaton
|
1f5c70da2e | chore: release channels v0.7.2 | ||
|
Tyler Slaton
|
f1b42e2e0a |
feat(skills): make setup-slack-channel drive the browser by default (refs OSS-705) (#6370)
## Summary Retesting the Slack setup end to end found the skill **too cautious to be useful**. It stopped at almost every step, where the earlier version stopped only for passwords and got through — and the run was rescued by the developer telling the agent outright to just control their browser, which cut human involvement down to typing passwords. A run that pauses at every control is slower than the manual path it replaced. Three changes, all pulling the same direction. ## 1. Driving is now the default, not implicit The skill said *"most of this workflow happens in a browser"* — descriptive, and it never told the agent to **drive** that browser. It now does, explicitly, and checks its own capability before Phase 0 rather than assuming either way. ## 2. When there is no browser, ask for one — per harness Generic advice is useless here, because enabling browser control differs by harness. The agent now works out which one it is in and names the single applicable route: Claude Code and Codex each ship their own support and enable it differently, most other harnesses take a general browser-use MCP server such as Playwright MCP. **If it isn't sure, it looks it up rather than guessing.** It also names the payoff — driving turns this into typing three secrets, the fallback is roughly fifteen manual browser steps — because that is what turns a shrug into a yes. The step-by-step walkthrough survives only for an explicit decline. ## 3. Consent is batched into one authorization Phase 0 now takes **one** yes naming the whole sequence: the Slack app from the wizard manifest, its install into a named workspace, the Channel, the adapter attach, and the project key. Phase 3 and `references/intelligence-channel.md` previously required *"state what you are about to change, get an explicit yes"* for **every** dashboard goal — and the reference said it twice, back to back. That is the concrete source of the stop-at-every-step behaviour. Reading the page before acting stays required; it is no longer a reason to check in. ## What did not change The secret boundary. The developer still types the bot token, signing secret, and API key themselves, and those remain the only mandatory stops alongside anything the authorization did not cover. Batching consent must not batch away a password — that is called out in the text. `version` bumped 1.0.0 → 1.1.0. ## Validation - `tsx scripts/sync-plugin-skills.ts --check` → **plugin skill mirror in sync** - `scripts/__tests__/sync-plugin-skills.test.ts` → 10 tests passed - `prettier --check` clean on both changed files ## Companion The same shift landed on the hosted big prompt in [CopilotKit/website#445](https://github.com/CopilotKit/website/pull/445) — default to driving, one batched authorization, harness-aware capability request. Both surfaces now say the same thing, which was the point of keeping them in step. Worth noting for reviewers: this takes Atai's side on *"where the skill asks permission to proceed, just do it"*, which was only half-applied before. It also reverses my own per-step confirmations from the first pass on #445 — the dogfooding run is the reason. |
||
|
Jerel John Velarde
|
7c5468e5d1 |
fix(skills): ask for the decisions a driving agent must not invent
Batching the Phase 0 authorization removed the per-goal confirmations that were
accidentally serving as decision points. Nothing then asked the developer for the
inputs the agent cannot legitimately choose, and Phase 1 still said "Enter a
Display name" in the imperative — so an autonomous run named the bot itself.
That name is the expensive one. The wizard derives the Channel Code from it, the
Code is what createChannel({ name }) declares and what the developer types as
/invite, and Slack bot names are workspace-wide — Phase 1 already warns that a
collision blocks the install. An agent that settles it has named someone's bot for
them and can fail the install doing it.
Phase 0 now gathers four decisions in one exchange before any browser opens: the
display name, the workspace, the test channel, and whether this is throwaway.
Phase 1 consumes the chosen name instead of inventing one, Phase 1's workspace step
uses the named workspace, and Phase 2's invite names the agreed channel and says
the developer runs it.
States the rule the whole design turns on: the decisions are inputs you cannot
invent, the authorization is permission you need once, and collapsing the second
does not license skipping the first.
|
||
|
Tyler Slaton
|
7193856d96 |
fix(runtime): ignore late lock renewal failures (#6369)
## Summary - stop handling in-flight lock renewal failures after the run has settled - keep aborting when a renewal fails during an active run - cover the completion-before-renewal race with a regression test ## Why Intelligence releases the thread lock after it accepts a terminal run event. A renewal that was already in flight can then return a 409. Clearing the interval stops future renewals, but it does not cancel that pending promise, so Runtime logged an error and called `abortRun()` after the run had completed. The lifecycle guard makes that late rejection a no-op. Active-run renewal failures still follow the existing abort path. ## Testing - `pnpm nx test @copilotkit/runtime` — 1,866 tests passed - `pnpm nx run @copilotkit/runtime:check-types` - `pnpm nx build @copilotkit/runtime` - `pnpm exec oxlint packages/runtime/src/v2/runtime/handlers/intelligence/run.ts packages/runtime/src/v2/runtime/__tests__/intelligence-lock-heartbeat.test.ts` - `pnpm exec oxfmt --check packages/runtime/src/v2/runtime/handlers/intelligence/run.ts packages/runtime/src/v2/runtime/__tests__/intelligence-lock-heartbeat.test.ts` The pre-commit hook also passed affected tests, `publint`, and `attw`. The repo-wide `pnpm check-format` still reports 25 unrelated files already present on `main`; both changed files pass the focused format check. |
||
|
Jerel John Velarde
|
ae0dbc2fd7 |
feat(skills): make setup-slack-channel drive the browser by default
Retesting the Slack setup end to end found the skill too cautious to be useful. It stopped at almost every step, where the earlier version stopped only for passwords and got through — and the run was rescued by the developer telling the agent outright to just control their browser. A run that pauses at every control is slower than the manual path it replaced. Three changes, all pulling the same direction. Driving the browser is now stated as the default rather than left implicit in "most of this workflow happens in a browser". When the agent has no browser tool it asks the developer to install one before starting, and names the route for the harness it is actually running in — Claude Code and Codex enable this differently, most other harnesses want a browser-use MCP server — with an instruction to look it up rather than guess. The manual walkthrough survives only for an explicit decline. Consent is batched into one Phase 0 authorization naming the whole sequence: the Slack app from the wizard manifest, its install, the Channel, the adapter attach, and the API key. Phase 3 and the Intelligence reference previously required "state what you are about to change, get an explicit yes" for every dashboard goal, and the reference said it twice. Reading the page before acting stays required; it is no longer a reason to stop. The secret boundary is untouched: the developer still types the bot token, signing secret, and API key themselves, and those remain the only mandatory stops alongside anything the authorization did not cover. |
||
|
David Sanchez
|
22108c0948 |
docs(pydantic-ai): constrain the direct dep, not the transitive one
Follows the maintainer's Correction #2 on issue 6363. An exact version in a docs install command is the same rot as the starlette==0.45.3 pin it replaced: it goes stale silently and nobody re-resolves prose. The 2.22.0 the docs shipped was already a version behind current the day it was written. - docs install lines use pydantic-ai-slim[ag-ui,openai]>=2,<3, which constrains the dep the pages actually care about and fails loudly at the v3 boundary - ag-ui-protocol drops out of the docs lines entirely; no doc snippet imports ag_ui, so naming it there was the transitive-dep noise the correction is about - starlette>=0.46.2 stays, because the v2 snippets import Starlette directly. A floor with no ceiling cannot force a downgrade, so it does not recreate the silent backtrack - examples/showcases/pydantic-ai-todos moves to a range in pyproject.toml and relocks; the uv.lock is what reproduces - examples/canvas/pydantic-ai keeps exact pins: it has no lockfile, so requirements.txt is its only reproducibility artifact Smoke-tested the open question from the issue: starlette 1.x works on pydantic-ai v2. All 8 doc pages pass on 2.23.0 + starlette 1.3.1 and on 2.23.0 + starlette 0.52.1, so Jordan's <1.0 guard can be dropped rather than raised. |
||
|
Tyler Slaton
|
86cd674c7c | fix(runtime): ignore late lock renewal failures | ||
|
Ben Taylor
|
8e59bfd16b |
feat(skills): add channels-setup pointing at the hosted onboarding guide (#6366)
## Summary The Channels onboarding workflow is served at **https://copilotkit.ai/channels-guide.md**, and every other entry point now copies one line that points there — the docs surfaces in #6357, the website strip in [website#444](https://github.com/CopilotKit/website/pull/444), the README in [channels-sdk#15](https://github.com/CopilotKit/channels-sdk/pull/15). Coding agents reached through a **skill** had no such pointer. They matched `setup-slack-channel`, which is scoped to Slack, to the provider half, and to an OpenTag checkout — so "help me get my agent into Teams" landed on a workflow that does not cover it. `channels-setup` closes that gap as a **pointer, not a copy**. The workflow stays in one place and is corrected there, instead of becoming a seventh surface that drifts against the CLI on its own schedule. ## Verifying the fetch is the substance of the file A thin pointer has one non-obvious failure mode, and it is the reason this skill is more than two sentences: ``` $ curl -sL -o /dev/null -w "%{http_code} %{content_type}\n" https://copilotkit.ai/channels-guide.md 200 text/html; charset=utf-8 $ curl -sL https://www.copilotkit.ai/channels-guide.md | grep -o "<title>[^<]*</title>" <title>Page not found | CopilotKit | ...</title> ``` **A missing guide does not return 404.** The site answers unknown paths with a "Page not found" HTML page under HTTP 200. An agent that keys on the status code gets a 73KB marketing page, concludes the fetch succeeded, and improvises channel setup from memory — which is exactly what the guide's boundaries exist to prevent, and it fails in the most expensive way available here: the project installs cleanly and answers nothing. So the skill checks the **body**: markdown rather than HTML, the guide's `# Build and prove a CopilotKit Channels agent` H1, and five `## Phase` headings. If any fail it stops and hands the user the URL, no matter what the status code said. ## Notes - **`RESERVED_LIFECYCLE_SLUGS` is not optional.** Standalone skills are not generated from `packages/*/skills`, so without the entry `sync-plugin-skills` treats the directory as an orphan and deletes it. Test updated alongside, including the hard-coded `size` (10 → 11). - **No existing skill is modified.** `setup-slack-channel` and `copilotkit-channels` keep their descriptions and bodies. Three skills now match channel work; if that proves too ambiguous in practice, narrowing the other two's `description` frontmatter is a follow-up, not a body rewrite. - **No manifest change.** `plugin.json` and `marketplace.json` do not enumerate skills individually. - No changeset — `packages/**` is untouched. ## Blocked on **The guide is not live yet.** [website#444](https://github.com/CopilotKit/website/pull/444) is still open, so the URL currently serves the 404 page shown above. This skill is inert until that merges and deploys — at which point it works with no further change here. The sentinel means the pre-deploy state is a clean stop rather than a wrong answer, so merging early is safe; it just isn't useful yet. ## Validation - `npx vitest run scripts/__tests__/sync-plugin-skills.test.ts` — 10 passed - `pnpm run check:plugin-skills` — `plugin skill mirror in sync`, exit 0; `skills/channels-setup/` survives the orphan pass - `npx oxfmt --check` on both changed TS files — clean - `npx oxlint scripts/` — 0 errors - `SKILL.md` frontmatter parsed with js-yaml: `name=channels-setup`, `version=1.0.0`, description 660 chars - Soft-404 behaviour confirmed against production with the `curl` commands above 🤖 Generated with [Claude Code](https://claude.com/claude-code) |
||
|
Mark
|
14a294f2d8 |
Port the Pydantic AI example to Pydantic AI v2 (#6355)
This pull request was posted by Claude Code using claude-opus-5 on behalf of David. David has not reviewed this diff. `examples/integrations/pydantic-ai` only runs on Pydantic AI **v1**. `Agent.to_ag_ui()`, `AGUIApp`, and the `pydantic_ai.ag_ui` shim were all removed in v2 (pydantic/pydantic-ai#5464, announced in pydantic/pydantic-ai#5345). The example pins `pydantic-ai-slim==1.0.18`, so anyone installing it against current Pydantic AI (2.22.0) fails at import. This ports it to v2. ## Changes 1. `agent/pyproject.toml` — `pydantic-ai-slim[ag-ui,openai]>=2.0.0`, `ag-ui-protocol>=0.1.19` 2. `agent/src/agent.py` — `StateDeps` moved from `pydantic_ai.ag_ui` to `pydantic_ai.ui` 3. `agent/src/main.py` — serve via `AGUIAdapter.dispatch_request` on a Starlette route 4. `agent/uv.lock` — relocked (resolves `pydantic-ai-slim` 2.22.0, `ag-ui-protocol` 0.1.19) ## One fix beyond the mechanical port The old wiring built the app once around a single shared `StateDeps` instance. `dispatch_request` mutates `deps.state` with the state the client sends, so one shared instance lets state leak between threads, channels and users — which matters more for Channels than it did for a single browser tab. Each request now gets its own `replace(deps)` copy, matching the pattern in [Pydantic AI's own AG-UI examples](https://github.com/pydantic/pydantic-ai/blob/main/examples/pydantic_ai_examples/ag_ui/api/shared_state.py). <details><summary>Verified end to end</summary> `uv sync` + a request through the actual ASGI app (model overridden with `TestModel` so no API call), with a Channels-shaped payload (`threadId`, `state`, `forwardedProps`): ``` health: 200 {'status': 'ok'} POST / -> 200 text/event-stream; charset=utf-8 event types: ['RUN_STARTED', 'TOOL_CALL_START', 'TOOL_CALL_END', 'TOOL_CALL_START', 'TOOL_CALL_ARGS', 'TOOL_CALL_END', ..., 'TOOL_CALL_RESULT', 'STATE_SNAPSHOT', 'TOOL_CALL_RESULT', 'STATE_SNAPSHOT', 'TEXT_MESSAGE_START', 'TEXT_MESSAGE_CONTENT', ..., 'TEXT_MESSAGE_END', 'RUN_FINISHED'] threadId echoed: slack-C123-thread-1 tool call: get_proverbs tool call: add_proverbs tool call: set_proverbs tool call: get_weather ``` The emitted event set is exactly what `channels-slack`'s `RunRenderer` subscribes to (`RUN_*`, `TEXT_MESSAGE_*`, `TOOL_CALL_*`), plus `STATE_SNAPSHOT` which it ignores. </details> <details><summary>Note on <code>ag-ui-protocol</code>: 0.1.19, not 0.1.18</summary> Not required by this port, but worth pinning forward: typed multimodal input content (`ImageInputContent` &c.) landed in 0.1.15 and the interrupt lifecycle in 0.1.19. We found that a Pydantic AI install below 0.1.15 rejects an inbound image attachment with a 422 rather than skipping it — so a Channels gateway forwarding a Slack image needs the newer floor. We're tracking that on our side. </details> |
||
|
David Sanchez
|
d19bd0f81c |
docs(pydantic-ai): port integration docs and demos to Pydantic AI v2
Agent.to_ag_ui(), AGUIApp and the pydantic_ai.ag_ui module were removed in Pydantic AI v2. The docs installed pydantic-ai unpinned, so anyone following the quickstart got 2.22.0 and failed first at dependency resolution (starlette==0.45.3 conflicts with the >=0.46.2 the ag-ui extra needs) and then at AttributeError. - 8 doc pages under showcase/shell-docs .../integrations/pydantic-ai serve the agent from a Starlette route via AGUIAdapter.dispatch_request - StateDeps moves from pydantic_ai.ag_ui to pydantic_ai.ui - stateful snippets build StateDeps per request; dispatch_request writes the client's state into deps.state, so a shared instance leaks state between users - install commands exact-pin pydantic-ai-slim==2.22.0 and ag-ui-protocol==0.1.19 - examples/canvas/pydantic-ai and examples/showcases/pydantic-ai-todos ported and pinned, todos relocked - skills/copilotkit-integrations reference updated to the same shape showcase/integrations/pydantic-ai is deliberately untouched; it is tracked separately. |
||
|
Alem Tuzlak
|
7b5cc8ccf9 |
fix(channels): recover from transient gateway outages (#6347)
## What changed - Mark initial gateway HTTP 5xx and transient transport failures as retryable. - Retry initial managed Channel activation with exponential backoff from 1 second to a 30-second cap until it connects or the manager stops. - Preserve retry hints from `gateway_draining` join replies and retry initial join timeouts. - Keep HTTP 4xx and NXDOMAIN failures terminal. - Back off established-session outage reminders from 30 seconds to a 15-minute cap while Phoenix continues reconnecting. ## Why The OpenTag Railway runtime saw the gateway host return HTTP 502 during an outage. Established Phoenix sessions keep retrying, but a runtime that starts during the outage stops after its one initial connect window. It cannot recover when the gateway comes back unless the process restarts. Fixed 30-second reminder logs also flood long outages. The gateway drain work now rejects new joins with a structured retryable response. The client must preserve that response so the runtime can retry instead of leaving the Channel in a terminal error state. ## Companion change CopilotKit/OpenTag#25 keeps the Railway HTTP server alive while an initial Channel retry is pending. OpenTag must consume a CopilotKit release containing this PR before that companion change can recover by itself. ## Validation - `pnpm nx run-many -t test,check-types,build -p @copilotkit/runtime,@copilotkit/channels-intelligence` - `pnpm nx run-many -t publint,attw -p @copilotkit/runtime,@copilotkit/channels-intelligence` - pre-commit tests and package checks for all affected projects - `pnpm exec oxfmt --check` on all five changed files - `pnpm exec oxlint` on all five changed files - `git diff --check` |
||
|
David Sanchez
|
4be25161b1 |
fix(examples): pin pydantic-ai starter deps and build deps per request
Address review on CopilotKit/CopilotKit#6355: - Pin `pydantic-ai-slim[ag-ui,openai]==2.22.0` and `ag-ui-protocol==0.1.19`, matching the starter fleet standard. Open-ended floors would pull a breaking major on release, and `docker/Dockerfile.agent:16` runs a bare `uv sync` that would silently re-resolve forward. - Mirror the specifiers in `uv.lock`'s `requires-dist` so `uv sync --frozen` at `Dockerfile:42` keeps working. Resolved versions and hashes are unchanged; `uv lock` is a no-op. - Construct `StateDeps(ProverbsState())` per request instead of `dataclasses.replace()`. `replace()` is a shallow copy, so the new deps point at the same state object; it is only safe today because `StateDeps` has one field that the adapter rebinds before every run. Any deps class with a second mutable field would silently share it. |
||
|
Benjamin Taylor
|
4c2ba2d161 |
feat(skills): add channels-setup pointing at the hosted onboarding guide
The Channels onboarding workflow is served at https://copilotkit.ai/channels-guide.md, and every other entry point now copies one line that points there (#6357 for the docs surfaces). Coding agents invoked through a skill had no such pointer, so they fell back to setup-slack-channel, which is scoped to Slack, to the provider half, and to an OpenTag checkout. channels-setup is a pointer, not a copy. The workflow stays in one place and is corrected there instead of drifting against the CLI on its own schedule. Verifying the fetch is the substance of the file. The site answers unknown paths with a "Page not found" HTML page under HTTP 200, so a status check proves nothing: an agent that keys on the status code gets a marketing 404 and improvises the workflow from memory. The skill checks the body instead -- markdown, the guide's H1, five Phase headings -- and stops if any fail. Standalone skills are not generated from packages/*/skills, so the slug is registered in RESERVED_LIFECYCLE_SLUGS. Without that entry the sync treats the directory as an orphan and deletes it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
Ben Taylor
|
397c8ce2e4 |
docs: route every Channels entry point at one onboarding skill (#6357)
## Visual Docs `/slack` — the featured panel from #6356, now carrying the one-line pointer. | Desktop screenshot | Mobile screenshot | |---|---| |  |  | | Desktop scroll | Mobile scroll | |---|---| | <video src="https://github.com/CopilotKit/website/releases/download/channels-entry-points-2026-08-03/docs-slack-desktop-scroll-7f14af56.webm" controls width="400"></video> | <video src="https://github.com/CopilotKit/website/releases/download/channels-entry-points-2026-08-03/docs-slack-mobile-scroll-7f14af56.webm" controls width="240"></video> | Docs `/teams` — same component, Teams wording, proving the frontend switch. | Desktop screenshot | Desktop scroll | |---|---| |  |  | <sub>Artifacts are hosted on a CopilotKit/website release tag rather than a new tag in this repo, to keep clear of the monorepo's release automation.</sub> --- ## Summary Builds on #6356 (merged in here, its featured-panel treatment kept) and finishes the job it started. #6356 made the Channels starter prompts **discoverable**. This makes them **correct**, and makes every other road say the same thing. The onboarding workflow was written out as prose on six surfaces across three repos: the Channels overview page (Slack and Teams), the docs landing activation strip, copilotkit.ai/channels, and the channels-sdk README twice. They drifted apart, and each went stale against the CLI on its own schedule. That drift is what the dogfooding feedback was reporting: - **"It didn't tell me which skills to install, so it dropped me into a list of god knows how many."** Neither copy button named a skill. - **"`npx copilotkit project select` didn't work — I had an old CLI."** Only some surfaces pinned `@latest`. - **An agent halted on `copilotkit channels`.** No such command exists in the published CLI — verified against 4.5.1, whose `--help` has no `channels` entry. - **"Copy this prompt is super easy to miss."** The strongest prompts were behind an accordion; the two visible buttons had invisible payloads. ## What changed - **`<ChannelsStartPrompt />`** — one shared entry point, frontend-aware, wearing #6356's featured treatment: accent panel, terminal mark, eyebrow, prominent copy action. - **Both overview accordions are gone.** The accordion existed because the payload was twenty lines. The payload is now one sentence, so hiding it behind "Open & copy prompt" costs a click and buys nothing. - **The activation strip emits the same pointer**, rendered on screen rather than living only in a clipboard payload. - **Both surfaces send `promptCopied` with a `surface` property**, so the funnel can answer which road people take. The website emits its own event name with the same property for the same reason. The prompt every surface now emits: ``` Run `npx copilotkit@latest skills install --skill setup-slack-channel -y`, then follow that skill to build your first CopilotKit Channels agent and connect it to Slack using CopilotKit's built-in agent. ``` ## Notes for reviewers - **`-y` respects `--skill`** — verified empirically against `copilotkit@4.5.1`: exactly one skill installs, no picker. The installer detects the coding agent it runs inside, so `--agent` is omitted. - **The panel renders exactly what the button copies.** Two earlier shapes were wrong in instructive ways: a full monospace paragraph wrapped like a rendering bug, and a code block with the ask underneath read as a shell command with a footnote — which made a button labelled "Copy prompt" look like it was lying. - **The `featured` Accordion variant from #6356 stays in `mdx-components.tsx`** as a shared opt-in capability, but its only two consumers were the accordions removed here, so it is currently unused. Worth a call: keep it, or drop it in a follow-up. - **`setup-slack-channel` is the slug the combined onboarding prompt will occupy** once it lands from channels-sdk. Its frontmatter is currently scoped hard to the Slack provider half, so it needs rescoping before this reads correctly for Teams. Companion PRs: CopilotKit/website#444 and the channels-sdk README. ## Validation - `npx vitest run` in `showcase/shell-docs` — 54 files, 372 tests passed - `npx tsc --noEmit` — clean - pre-commit lint and commitlint hooks passed - Rendered and reviewed `/slack` locally; confirmed the copied string matches the rendered one ## Update — impressions and a copy-only prompt Since the artifacts above were first posted: - **Impression event.** `docs.channels_activation_viewed` fires once per surface on first intersection at 50%. Both docs entry points previously emitted a copy event and nothing else, so the copy count had no denominator. `surface` values now live in a shared `CHANNELS_ACTIVATION_SURFACES` map, keeping the landing strip, the overview panel, and copilotkit.ai/channels separable inside one funnel. - **Guarded on `typeof IntersectionObserver`.** An impression is never worth breaking a render for, and this repo's jsdom tests do not define it. - **The prompt is no longer rendered.** Per review, the copy button carries the payload. The panel leads with the copy action directly under the heading — the prompt is not on the page, so the button is the point of the panel rather than trailing furniture — and the supporting line no longer says "paste this" next to nothing. Screenshots and videos above were re-captured after all of it. ## Validation (current) - `npx vitest run` in `showcase/shell-docs` — 55 files, 376 tests passed - `npx tsc --noEmit` — clean - pre-commit lint and commitlint hooks passed - Reviewed `/slack` and `/teams` at 1440×900 and iPhone 13 ## Companion PRs - CopilotKit/channels-sdk#15 — **merged** - CopilotKit/website#444 — open --- ## Update — addresses both review blockers ### Blocking issue: the skill did not support the promised workflow — resolved by removing the skill from the path The pointer is now a fetch of one hosted file, not a skill install: ``` Read https://copilotkit.ai/channels-guide.md and help the user build their first channel ``` The guide lives at `public/channels-guide.md` in CopilotKit/website and owns the whole workflow. It asks the developer which platform and which agent framework they want, which is why the pointer passes neither. That removes the mismatch at its root rather than narrowing the copy around it. Naming the picker's channel and backend meant these pages promised coverage on `setup-slack-channel`'s behalf, and that skill is scoped to Slack, to the provider half, and to an OpenTag checkout — so Teams pointed at a workflow that does not exist and the picker implied nineteen backends it never claimed. A pointer that names nothing cannot overpromise. `CHANNELS_ONBOARDING_SKILL`, its install command, and the per-selection prompt builders are gone; one constant serves every surface. ### Visual regression: disclosure restored, treatment re-tokenised | Collapsed | Expanded | |---|---| |  |  | | Mobile | Teams variant | |---|---| |  |  | - **The expandable preview is back.** The panel is the shared featured `<Accordion>` again — reused, not re-implemented — so the overview stays compact and a reader can expand to read the exact prompt before copying. - **Whitespace.** Collapsed height went from ~380px to ~110px with `p-4`, matching every other docs panel. The page's own introduction is visible without scrolling. - **Colour.** You were right, and my "it is unchanged from #6356" reply missed the point. `copilotkit-ui-theme` names *"purple accent bar or stripe"* as a known wrong direction, and `copilotkit-branding` requires accent to be restrained with gradients behind content rather than as the contrast layer — the saturated `--accent` tile plus accent-mixed gradient was both at once. It now matches the `OpsPlatformCTA` idiom: `--bg-elevated`, `--border`, `--shadow-control`, accent only on a 20px glyph and the hover state. Verified light and dark. One deliberate deviation: I dropped the "READY-TO-USE STARTER PROMPT" eyebrow on the low-noise rule, since the title says the same thing. Easy to restore if you want it. ### Behavioural bug: fixed `copyPrompt` shared one `try` with `posthog.capture`, so a throwing client reported "Copy blocked" for a prompt already on the clipboard. Capture now sits behind the same isolated helper `ChannelsActivationStrip` uses. Two regression tests: capture throwing after a resolved write still shows "Copied"; a rejected write still shows "Copy blocked" and emits no copy event. ### Lockfile churn: dropped Reverted to `origin/main`. It came from an `npm install` in the review worktree; no dependency changed. ### Also added `docs.channels_activation_prompt_expanded` — the disclosure is a funnel step neither `viewed` nor `promptCopied` can see, so someone who never opened the panel was indistinguishable from someone who opened it and left. ## Validation (current) - `npx vitest run` in `showcase/shell-docs` — 55 files, 376 tests passed - `npx tsc --noEmit` — clean; pre-commit lint and commitlint passed - Reviewed `/slack` and `/teams`, collapsed and expanded, light and dark, at 1440×900 and iPhone 13 |
||
|
Tyler Slaton
|
d92652dcdd |
chore: release monorepo v1.66.1 (#6365)
🚀 pkg-pr-new / build (push) Has been cancelled
## Release monorepo v1.66.1 **Scope:** `monorepo` | **Bump:** `patch` --- ### How this release process works 1. **This PR was created automatically** by the "release / create-pr" workflow. It bumped the `monorepo` packages to `1.66.1` and generated AI-enhanced release notes. 2. **CI runs on this PR** — the full test suite (unit tests, lint, type checks, build) must pass before merging. This is the review gate. 3. **Review the release notes** in `release-notes.md` in this PR. If a Notion draft was created, you can edit the release notes there before merging. 4. **When this PR is merged**, the `release / publish` workflow automatically: - Builds all packages - Publishes the `monorepo` packages to npm at version `1.66.1` - Creates git tag `monorepo/v1.66.1` - Creates a GitHub Release with the final release notes ### Before merging - [ ] CI is green (tests, lint, types, build) - [ ] Version bumps look correct - [ ] Release notes are accurate (edit in Notion if a draft was created) --- > **Do not merge until CI is fully green.** The full test suite runs automatically on this PR.v1.66.1 |
||
|
tylerslaton
|
c69f7e96a5 | chore: release monorepo v1.66.1 | ||
|
Benjamin Taylor
|
a1d27ab017 |
feat(shell-docs): strip both Channels surfaces to one copy action
**Root page.** The channel and agent-backend pickers are gone, and the copy action moves beside the supporting sentence, under the heading. The pickers had stopped earning their place: the guide asks which platform and framework the developer wants, so choosing here asked the same question twice and changed nothing about what got copied. `ActivationSelect` and its option plumbing go with them. The setup-guide link becomes one aside alongside OpenTag rather than a per-selection route. **Per-framework pages.** The accordion is gone. It existed to keep a twenty-line prompt out of the way; the payload is now a single action, so a disclosure cost a click and revealed nothing. The panel keeps the in-content idiom it shares with `OpsPlatformCTA` — neutral surface, `--border`, accent on a small glyph — and the button reads "Copy prompt" like every other surface. `docs.channels_activation_prompt_expanded` retires with the disclosure that was its only trigger. |
||
|
Tyler Slaton
|
d8eeb77aef |
chore: release channels v0.7.1 (#6361)
🚀 pkg-pr-new / build (push) Has been cancelled
## Release channels v0.7.1 **Scope:** `channels` | **Bump:** `patch` --- ### How this release process works 1. **This PR was created automatically** by the "release / create-pr" workflow. It bumped the `channels` packages to `0.7.1` and generated AI-enhanced release notes. 2. **CI runs on this PR** — the full test suite (unit tests, lint, type checks, build) must pass before merging. This is the review gate. 3. **Review the release notes** in `release-notes.md` in this PR. If a Notion draft was created, you can edit the release notes there before merging. 4. **When this PR is merged**, the `release / publish` workflow automatically: - Builds all packages - Publishes the `channels` packages to npm at version `0.7.1` - Creates git tag `channels/v0.7.1` - Creates a GitHub Release with the final release notes ### Before merging - [ ] CI is green (tests, lint, types, build) - [ ] Version bumps look correct - [ ] Release notes are accurate (edit in Notion if a draft was created) --- > **Do not merge until CI is fully green.** The full test suite runs automatically on this PR.channels/v0.7.1 |
||
|
tylerslaton
|
f8f9e2a721 | chore: release channels v0.7.1 | ||
|
Tyler Slaton
|
dfe728d833 | fix(channels): retry gateway drain joins | ||
|
Tyler Slaton
|
82eaecdfda | fix(channels): retry transient gateway activation | ||
|
Tyler Slaton
|
503fc7c593 |
fix(channels): back off prolonged outage logs
Keep drop, give-up, and recovery logs immediate. Space repeated still-down reminders during long gateway outages. |
||
|
Benjamin Taylor
|
335b12a96a |
feat(shell-docs): point the Channels prompt at the hosted guide
Replaces the skill-install pointer with a fetch of one hosted file:
Read https://copilotkit.ai/channels-guide.md and help the user build
their first channel
The guide lives at `public/channels-guide.md` on the marketing site and owns the
whole workflow. It asks the developer which platform and which agent framework
they want, which is why this pointer passes neither.
That is what resolves the review's blocking issue rather than papering over it.
Interpolating the picker's channel and backend meant these pages promised
coverage on a skill's behalf — and the skill it named is scoped to Slack, to the
provider half, and to an OpenTag checkout, so the Teams road pointed at a
workflow that does not exist and the backend picker implied nineteen it never
claimed. A pointer that names nothing cannot overpromise, and the guide handles
selection itself.
Consequently `CHANNELS_ONBOARDING_SKILL`, its install command, and the
per-selection prompt builders are gone: one constant serves every surface.
Supporting copy that said the prompt "installs the onboarding skill" or offered
a "tailored" prompt is corrected — neither is true now.
|
||
|
Benjamin Taylor
|
afef22d316 |
feat(shell-docs): restore the prompt disclosure and put it on the docs palette
Two changes from review. **The disclosure is back.** The panel is the shared featured `<Accordion>` again — reused, not restyled — so the overview stays compact when collapsed and a reader can expand to read the exact prompt before copying it. Copy-only made the payload opaque, which was the problem #6356 set out to fix. The component still owns the Slack/Teams switch and the analytics; the container is markup. **The treatment is now token-only.** The featured variant carried a saturated `--accent` tile and an accent-mixed gradient. `copilotkit-ui-theme` names a purple accent bar or stripe as a known wrong direction, and `copilotkit-branding` scopes accent to restrained, atmospheric use with gradients behind content rather than as the contrast layer — the old treatment was both at once, on a docs `--accent` that resolves to violet. It now matches the in-content panel idiom already in `OpsPlatformCTA`: `--bg-elevated`, `--border`, `--shadow-control`, and accent carried only by a small glyph and the hover state. Padding drops to `p-4` like every other docs panel, so a collapsed prompt no longer pushes the page's own introduction below the fold. Verified in light and dark. Also adds `docs.channels_activation_prompt_expanded`. The disclosure is where a funnel loses people and neither `viewed` nor `promptCopied` can see it: someone who never opened the panel is indistinguishable from someone who opened it and walked away. The prompt inside wraps instead of scrolling. The docs' usual code block scrolls horizontally, which is right for code and wrong here — it hid half the prompt behind the overflow, defeating the point of letting people read it first. |
||
|
Maxim
|
bc80efb17f |
Reskinnable demo: a health-system knowledge desk with grounded retrieval and long-running approval gates (#6321)
Adds a skin to `examples/showcases/reskinnable-demo`: **Keel** (`/keel`), the internal knowledge and operations desk of Harbor Point Health, a fictional nine-hospital health system. The product is one sentence: **an answer you can trust, that turns into work that actually happens.** ## Why this skin The shipped skins bracket the `Skin` contract — `airline` is in-memory with four optional slots, `banking` is REST-backed with nearly all of them, and #6302's `logistics` proves a server-enforced authority gate. A further skin earns its place only by demonstrating something none of them do. Keel introduces **two substrates the app has never had**: 1. **Grounded retrieval.** A server-side `search_knowledge` tool over a nine-document policy corpus, returning cited passages. Every existing tool in this app pushes client state *up* to the agent; this inverts the flow — the agent reaches for something the user has not seen and brings it back. 2. **Long-running stateful process.** Runs whose steps advance on a 900ms ticker and halt at human approval gates. Banking's HITL is a single approval; airline's mutations are instantaneous. Neither has duration — state that changes while you are not looking. It is also the first skin with **parameterized routes** (`knowledge/<docId>`, `runs/<runId>`), so `resolvePage` actually destructures the segment array rather than using a flat lookup. ## The demo arc | Beat | What happens | | --- | --- | | *"What's our policy on giving a contractor access to patient records?"* | `search_knowledge` → grounded answer + `showSources` citing POL-114 and POL-302 | | Click a citation | Routes to `/keel/knowledge/phi-access-policy#minimum-necessary`, scrolls, highlights | | *"Set that up for Priya Raman"* | `showPlaybook` → `startRun` HITL plan preview (6 steps · 1 gate · 5 roles) | | Run ticks | `RUN-1045` advances, then halts **blocked** on Privacy Officer sign-off | | Switch persona → *"approve it"* | `approveStep` card carrying the policy citation as the reason the gate exists | **The citation resolving into the real app is the load-bearing beat.** Most AI-knowledge demos show a footnote; few make the footnote drive the app. ## Two details that are load-bearing, not stylistic **`showSources` cannot render a fabricated citation.** It accepts only `(docId, sectionId)` pairs — never citation text — and resolves the ref, heading, and snippet from the real corpus client-side. A citation the model invents fails to resolve and is dropped, rather than rendering as a convincing fake. Grounding is structural here, not prompt-dependent. **The ticker cannot thrash the agent's context.** A 900ms timer plus a `useAgentContext` readable is a token-cost hazard. The readables memoize on a `summaryKey` derived only from `(runId, status, currentStepId)` tuples — never from elapsed time — so context is rewritten on meaningful transitions and never on a raw tick. The interval is also cleared whenever nothing is running. ## What it exercises - **In-memory substrate** — seed-backed `useData` with a pure reducer (`tick` / `startRun` / `approveStep` / `rejectStep` / `cancelRun`) and a stale-approval race guard. - **Identity triad** — `RuntimeProviders` (persona above the provider) + `useRuntimeProperties` + a server-safe `identifyUser`, scoping memory per persona. A gate is actionable only when its `approverRole` matches the active persona; otherwise it renders "waiting on <role>". - **a2ui canvas** — an ops report from a deterministic op-builder; no figures travel in the ops, every number binds client-side. - **OGUI sandbox** — four functions projecting to allowlisted DTOs. - **Theme** — deep pine on warm neutrals at `--radius: 0.5rem`, the tightest in the app, with `--brand-violet` re-valued as the amber awaiting-approval accent (the shared vocabulary has no warning token, and a skin may not invent token names). Pure token re-valuing; no component edits, no `globals.css` change. Dark-capable via the `--nw-dark-capable` opt-in. Deliberately omits `chatHeaderActions` and `onSuggestionSelect` (they exist in banking only to serve its PDF-attachment beat) and `Providers` (nothing needs to mount below the provider). ## The most valuable file in the diff `src/skins/keel/data/seed.test.ts` asserts that **every one of the 23 playbook-step `policyRef`s resolves to a real `docId` + `sectionId`** in the corpus. That invariant is what fuses the two substrates — a dangling reference means an approval card cites a document that does not exist. It has already earned its keep: it caught a citation to `POL-208 §documentation` that did not exist (`documentation` belongs to POL-121), which had been written into the design spec and propagated into the implementation plan before any code was generated. That test was also **mutation-tested** rather than trusted: deliberately breaking one `policyRef` turns it red, so its green result reflects a real check and not one of its fallback paths silently no-opping. ## Verification Keel's own suite: **16 test files / 105 tests** (`npx vitest run src/skins/keel`). Whole app after merging `main`: **38 files / 215 tests** — the larger number includes logistics' 65 tests, which arrived with the merge and are not this PR's. `pnpm lint` 0 issues · `npx tsc --noEmit -p .` exit 0 · `pnpm build` compiles. `main` is merged in (resolving the append-only registry conflict predicted below), so the branch is mergeable and all four skins — banking, airline, logistics, keel — are registered under the same id in both the client and server registries. **No CI job exercises this app, so green checks here are not evidence.** `.github/workflows/static_quality.yml` carries `paths-ignore: ["examples/**"]` on both its `push` and `pull_request` triggers, and no workflow in the repo references `reskinnable-demo` at all. The passing checks on this PR are unrelated legacy example demos plus binary/config checks. The local gates above are the only real signal — worth knowing before reading a green tick as coverage. **Not verified: the chat agent itself.** No API key was available, so gen-UI, both HITL cards, the citation deep-link, the ticker's live advance, persona switching, dark mode, and OGUI are unexercised at runtime. A live smoke test of those paths is the main thing a reviewer should add — it is also where the `pointer-events-auto` fix and the context-churn guard would actually be observed working. ## Review pass The skin went through a four-round unbiased review (16 reviewers per round, non-leading prompt) plus a fix cycle per finding. Findings landed as six thematic commits on top of the feature commits: | Commit | Closes | | --- | --- | | `fix(keel): make engine mutations honest and single-path` | Lost updates from non-functional `setRuns`, a stale render-time read, gates that could enter `awaiting_approval` without `startedAt`, a rejector mislabelled as approver, and a mutation reporting success for a commit that did not happen | | `fix(keel): make retrieval deterministic and citations grounded` | Locale-dependent tiebreak ordering (breaking the documented byte-identical guarantee), order-dependent transitive synonym expansion, phrase synonyms matching across word boundaries, duplicate citation keys, and a second citation into the same document silently failing to re-scroll | | `fix(keel): guard prototype access on untrusted lookup keys` | `/keel/constructor` and friends resolving `Object.prototype` members past a `?? null` guard and being handed to React as a component; same class in the persona lookup, where the key arrives from client-forwarded properties | | `fix(keel): pin date/time formatting and fix in-chat interactivity` | Hydration mismatches from ambient-locale formatting, and policy links dead in chat because CopilotKit paints `useComponent` renders with `pointer-events: none` | | `fix(keel): anchor seeded runs to a per-call now` | Seed timestamps anchored to fixed calendar dates while the app compares them to `Date.now()` — a ~6.7-day cycle time poisoning the median KPI, and the mid-flight ticking run arriving already resolved on remount | | `fix(keel): harden report surface, catalog, and remaining chrome` | Duplicate a2ui component ids colliding React keys, and a failed report surface latching so the canvas never self-healed | Each fix carries a red-green proof (test written, source broken to confirm red, restored to confirm green). Two were structural rather than per-instance: the prototype-safe lookup uses a `Map` so inherited keys cannot exist, and a single `ChatSurface` wrapper now carries `pointer-events-auto` so exactly one such literal remains in the skin. **Caveat worth stating plainly: the final six fixes did not themselves get an unbiased review round.** The review was stopped deliberately after round four. Earlier rounds are the calibration — round one's 16 fixes introduced 3 regressions (~19%), round three's introduced zero — so the risk is real but bounded, and it is concentrated in these six commits rather than spread across the diff. ## Relationship to #6302 **Resolved: #6302 (`logistics` / Meridian) has merged, and keel landed second.** The predicted conflict materialised exactly as described — both PRs append to the same two registry lines — and is resolved keep-both in the merge commit on this branch. Worth noting that until that merge, GitHub created *no* CI runs for this PR at all: it does not evaluate `pull_request` workflows for a branch whose mergeable state is `CONFLICTING`, because it cannot compute the merge ref those events run against. A conflicting PR therefore looks like "CI hasn't started" rather than "CI failed". Worth deciding before either merges whether these ship as two skins or Keel's retrieval half folds into Meridian. Keel adds grounded retrieval, duration, and parameterized routes; it also re-proves approval gates, maximal contract coverage, and an a2ui canvas that Meridian already demonstrates. Keel's canvas (`canvas-surface.tsx` + `ops-report.ts`) was deliberately built as cleanly separable scope for exactly this reason — nothing imports it except the agent and the skin object. ## Known follow-ups (non-blocking) - In the `adverse-event` playbook, both the `rca-decision` and `close` steps cite `POL-208 §root-cause-analysis`. POL-208 has no section that fits "Document and close" better, so this is the least-bad mapping rather than a clean one; a new corpus section would resolve it properly. - `seed.test.ts`'s live-corpus check retains two fallback paths (`catch { return }` and `if (!docs) return`) that would let it pass silently if the corpus import ever broke. It genuinely runs today (verified by mutation), but the escape hatches are a latent hazard worth removing now that the corpus is guaranteed present. 🤖 Generated with [Claude Code](https://claude.com/claude-code) https://claude.ai/code/session_01FkMi54kqfNwAoUB3M3CBp3 |
||
|
Benjamin Taylor
|
4d5fd3a25e |
fix(shell-docs): stop a throwing analytics client from faking a copy failure
The clipboard write and the capture call shared one try block, so a PostHog client that threw reported "Copy blocked" for a prompt already sitting on the clipboard. Only the write decides what the reader is told; capture moves behind the same isolated helper the activation strip already uses, and the impression observer uses it too instead of its own inline catch. Two regression tests: capture throwing after a resolved write still shows "Copied", and a rejected write still shows "Copy blocked" without emitting a copy event. Also drops 16 lines of `dev: true` lockfile churn picked up from an `npm install` in the review worktree — no dependency actually changed. |
||
|
Maxim
|
0b1b37e0af |
Merge origin/main into feat/reskinnable-demo-keel-skin
Resolves the append-only registry conflict this PR's description predicted: #6302 (logistics / Meridian) landed first and took the same two lines, so keel takes the trivial conflict. Both edits are additive, so the resolution is keep-both in both registries — logistics first (it landed first), keel second. Why the merge was necessary rather than optional: GitHub does not create pull_request workflow runs for a PR whose mergeable state is CONFLICTING, because it cannot compute the merge ref those events run against. The push of the review fixes therefore produced ZERO CI runs — verified over 12 minutes of polling the Actions API by head SHA, and corroborated by Actions being healthy repo-wide at the same time. Not a timing artifact, and not a path filter: test_e2e-legacy-v1 and auto_merge_showcases both trigger on pull_request with paths: examples/**, and synchronize is a default activity type. Scope of the drift: main is 127 commits ahead of the merge-base, but within reskinnable-demo it added only logistics' own files (48 skin + 17 API route + 2 reskin-skill docs) plus +7 append-only lines across exactly the two registry files. No globals.css change, no skin-contract.ts change, no skins-config.ts change — so the shared token vocabulary and the frozen Skin contract are untouched, and keel's isolation claim still holds against a four-skin registry. All four skins are now registered under the same id in both registries (banking, airline, logistics, keel), which is the invariant the two-registry split exists to maintain. Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FkMi54kqfNwAoUB3M3CBp3 |
||
|
Maxim
|
c01b637bdc |
fix(keel): harden report surface, catalog, and remaining chrome
Remaining fixes across the ops-report builder, the a2ui canvas surface, the catalog, agent wiring, suggestions, and the knowledge/playbooks pages: correct the report surface output and its catalog registration so the canvas renders the intended report, and tighten the surrounding chrome so these entry points behave consistently. Tests cover the report builder and canvas surface. Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FkMi54kqfNwAoUB3M3CBp3 |
||
|
Maxim
|
28adff5dbf |
fix(keel): anchor seeded runs to a per-call now
Seed data used a fixed or module-load timestamp, so seeded run timelines drifted out of a plausible window as time passed and could not be reproduced deterministically in tests. This anchors every seeded run relative to a `now` passed in at seed time, keeping the demo's timeline coherent whenever it is generated and making the seed testable. A test pins the relative anchoring. Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FkMi54kqfNwAoUB3M3CBp3 |
||
|
Maxim
|
5bbe483073 |
fix(keel): pin date/time formatting and fix in-chat interactivity
Dates and times were formatted with the runtime's ambient locale and timezone, so server and client rendered different strings and React hydration mismatched; several call sites also built ad-hoc formatters that drifted from one another. This centralizes formatting behind pinned locale/timezone formatters — one per format — so output is stable across environments. It also fixes the ChatSurface pointer-events boundary so interactive controls rendered inside the chat transcript (playbook, approval, and run cards) actually receive clicks. Tests cover the pinned formatters. Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01FkMi54kqfNwAoUB3M3CBp3 |
