mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-14 08:53:19 +01:00
f4337bb5fb
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work > - The chat-feature evals measure whether agents complete real chat requests through the product > - The chat catalog eval harness and its repairs live on the `chat-eval/master-candidate` branch, not on master > - Today's work made 139 of 242 chat evals pass in verified single-case live runs. Each product fix landed on master as its own pull request > - The eval-side changes (setups, graders, fault seams, evidence capture) still exist only on the candidate > - This pull request lands the harness on master. Master stays the source of truth for product behavior > - The benefit is one tree: evals run against master commits, and product fixes reach the evals as soon as they merge ## Linked Issues or Issue Description Refs #19441, Refs #19599, Refs #20270, Refs #20334, Refs #20342, Refs #20344, Refs #20354, Refs #20360, Refs #20373, Refs #20382, Refs #20409 **What existing behavior does this improve?** The chat catalog eval harness and the product it evaluates live on a long-running branch outside master. Eval fixes do not reach master, so master-based eval runs use old setups and graders. **Subsystem affected** `tests/runner-e2e` chat catalog harness, plus a small set of product changes that exist only on the candidate. **Proposed change** Land the harness on master. Every product fix from this eval work already merged to master as its own pull request. Conflicts keep master's product behavior. ## What Changed - Most of the diff (808 of 851 files) is under `tests/runner-e2e/`: chat-catalog domains, scenarios, descriptors, the TestOps manifest, private evidence tooling, and their tests. - Graders now check saved state, setup literals and rendered or clicked links. Many keyword and regex checks on free-form replies are removed. Each changed grader has a bumped `oracleVersion` and calibration tests (a valid reply, a reworded valid reply, and broken variants that must fail). - The shared link helper `hasLink` no longer treats a list link (for example `/COM/issues?view=mine`) as a link to every task without a URL key. - Setup fixes build real state through supported product routes: routine history, agent grants through the board grant route, budget holds with a real cap, and a storage-layer archive fault for `inbox-search.archive-partial-failure`. - Browser boundary journeys (sign-in, chat-disabled, onboarding) now write the standard pass evidence files. - Retargeted cases, approved by the maintainer: `agent-chat.chat-disabled` (Conference Room gate), `agents.restore-instructions` (no saved instruction history), `access.manage-membership` (member permission edits stay board-only). - The private evidence trusted files (`manual-private-handoff.ts`, `manual-private-retention.ts`, `manual-private-report-assets.ts`) are byte-identical to master. - Product files outside `tests/` (43 files, +1,328 −49) carry candidate-only changes that master does not yet have. Most are tests. The main code changes are in `request-human-input.ts`, `heartbeat-run-summary.ts`, `native-completion-feedback.ts`, `native-run-finalizer.ts`, `openapi.ts`, `secrets.ts` and `IssueChatLongThreadPerf.tsx`. ## Verification - `pnpm vitest run chat-catalog-domains chat-catalog-contract.test.ts chat-catalog-consistency.test.ts unit-launcher.test.mjs --testTimeout=180000` in `tests/runner-e2e`: 269 files pass, 3,550 tests pass. - `node --test chat-full-catalog-testops-manifest.test.mjs`: 10 pass. - `node tests/runner-e2e/source-sdk-public-inputs.mjs <HEAD> --check`: no change. - `tsc` for `server` and `tests/runner-e2e`: clean. - Live single-case verification: 139 of 242 cases passed on paid runs, and a person read the saved evidence for each. These are single-case runs, not one full-cohort run. ## Risks - Large diff. Most of it is test-only, but the 43 non-test files change product code paths. Review those files first. - Some graders are now less strict on reply wording. A person must read replies for meaning on those cases. - 42 cases need disposable outside accounts (Notion, Slack, GitHub, Daytona and others) and stay blocked. `access.invite-teammate` stays failing until local agent isolation lands, because the runner's durable state keeps a live invite token. - The paid workflow compares the trusted private-evidence files with master. They are unchanged, so this risk is low. ## Model Used Anthropic Claude Opus 5.5 (`claude-opus-5-5`), with tool use and code execution, in Claude Code. Subagents used the same model. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change (e.g. `docs/...`, `fix/...`) and contains no internal Paperclip ticket id or instance-derived details - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups - [x] I will address all Greptile and reviewer comments before requesting merge 🤖 Generated with [Claude Code](https://claude.com/claude-code) --------- Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com> Co-authored-by: Paperclip <noreply@paperclip.ing>
91 lines
4.4 KiB
YAML
91 lines
4.4 KiB
YAML
# yaml-language-server: $schema=./catalog.schema.json
|
|
# Planning scenarios; not registered executions or evidence of a pass.
|
|
schema_version: 1
|
|
feature: agent-chat
|
|
status: planned
|
|
scenarios:
|
|
- id: agent-chat.reopen-conversation
|
|
capabilities:
|
|
- agent-chat/discovery
|
|
- agent-chat/history
|
|
context: >-
|
|
The person previously chatted with Morgan; the conversation is still accessible and Chat
|
|
is enabled. Opening the page itself must not start work.
|
|
query: >-
|
|
Take me back to my conversation with Morgan.
|
|
expected_outcome: >-
|
|
The supported human navigation path opens the person's existing conversation, with the
|
|
right agent and retained history. No duplicate standing chat or provider execution is
|
|
created just by navigation. The agent does not enumerate another person's private chat
|
|
list to satisfy the request.
|
|
- id: agent-chat.remember-and-follow-up
|
|
capabilities:
|
|
- agent-chat/history
|
|
context: >-
|
|
The previous turn selected a blue design and contains a saved preview file; the browser
|
|
has since reloaded.
|
|
query: >-
|
|
Use the blue design we picked earlier, and show me that preview again.
|
|
expected_outcome: >-
|
|
The reply resolves the reference from persisted conversation history and returns the same
|
|
accessible preview. It preserves the selected design in any follow-up task. The agent asks
|
|
a focused clarification if the history does not establish the choice rather than
|
|
fabricating a prior agreement.
|
|
- id: agent-chat.track-delegated-work
|
|
capabilities:
|
|
- agent-chat/handoff
|
|
- tasks/create
|
|
context: >-
|
|
This conversation already created a website task assigned to Morgan, and it is still
|
|
running.
|
|
query: >-
|
|
How is the website coming along?
|
|
expected_outcome: >-
|
|
The agent reads the linked task and available execution evidence, gives its current status
|
|
and links to real output when present. It does not create another website task or convert
|
|
the standing chat into the execution task. Completion updates remain associated with the
|
|
original request.
|
|
- id: agent-chat.multiple-handoffs
|
|
capabilities:
|
|
- agent-chat/handoff
|
|
- agent-chat/history
|
|
context: >-
|
|
The conversation has one active website task and a separate completed logo task.
|
|
query: >-
|
|
Change the website headline to Launch sooner, and leave the logo alone.
|
|
expected_outcome: >-
|
|
The new requirement reaches the existing website task and its authorized worker once. The
|
|
logo task and delivered logo remain unchanged. Chat identifies the affected task and
|
|
distinguishes feedback saved for later from feedback consumed by an active run.
|
|
- id: agent-chat.conference-room
|
|
capabilities:
|
|
- agent-chat/conference-room
|
|
context: >-
|
|
The separate board conference-room experiment is enabled only in the supported
|
|
local-trusted deployment; ordinary Agent Chat is also available.
|
|
query: >-
|
|
Let us discuss the launch in the conference room.
|
|
expected_outcome: >-
|
|
The agent points the person to the gated board discussion surface and explains its actual
|
|
scope. The journey opens that surface only in the supported deployment and retains normal
|
|
company context. It does not promise ordinary standing-chat history or execution parity
|
|
where the board-chat protocol differs.
|
|
- id: agent-chat.chat-disabled
|
|
capabilities:
|
|
- agent-chat/discovery
|
|
- agent-chat/conference-room
|
|
- developer-labs/route-gates
|
|
context: >-
|
|
Agent Chat is always available. The separate Conference Room experiment is off on this
|
|
target, and its opt-in is not offered in settings, so the board discussion surface is
|
|
unavailable. The person follows an old Conference Room link.
|
|
query: >-
|
|
Open the conference room so I can talk with Morgan.
|
|
expected_outcome: >-
|
|
The person does not land on a broken route or a Conference Room that cannot work: the old
|
|
link opens a working company page and navigation does not offer the Conference Room. The
|
|
supported alternative, an ordinary chat with Morgan, opens from Chat. The board discussion
|
|
API refuses while the experiment is off, and nothing creates a task, chat or run that
|
|
claims to be the conference room. Turning the experiment on is an instance operator
|
|
decision; an ordinary agent cannot change the instance setting.
|