DottaandPaperclip
0d22e27dc4
## Thinking Path
> - Paperclip runs agents through a controlled execution boundary.
> - Native Runner starts ACPX providers through a TypeScript sidecar and
a Rust transport.
> - That transport hides arbitrary provider messages because they can
contain credentials.
> - Grok admission failures currently have no stable code. Other startup
failures can also lose their owning stage.
> - A real native startup failure therefore reports only
`classification=unclassified`, even when the ordinary subscription hello
test succeeds.
> - This pull request preserves fixed failure categories without
forwarding provider text or changing admission policy.
> - The benefit is a useful next diagnostic result before another
functional repair or retry.
## Linked Issues or Issue Description
Refs #17922 and #17942.
**What happened?**
A native Grok session failed before its first tool. Runner assets passed
verification. The Rust rejection discarded the cause because only closed
error codes may cross that boundary. The existing admission guard used
plain errors. Startup errors outside that guard could also lose their
stage.
**Expected behavior**
Report a fixed Grok admission reason, or a fixed startup stage when no
known error code exists. Keep credentials, provider messages, session
IDs, and paths out of the classification. Keep the failure and cleanup
effective.
**Steps to reproduce**
1. Have a Grok ACP fixture return a session without its native effort
acknowledgement.
2. Start it through the real ACPX client and pass the failure through
the sidecar classifier.
3. Before this change, the result is `acpx_sidecar_command_failed`. Rust
reduces that to `unclassified`.
4. The same loss occurs for injected sandbox, credential, and runtime
admission failures.
## What Changed
- Give Grok effort admission errors fixed codes for missing
acknowledgement, model or effort mismatch, session mismatch, conflicting
configuration, drift, and invalid protocol use.
- Record closed host startup stages in a private weak map. Preserve the
original error object, cancellation, and resource cleanup.
- Prefer existing known error codes over a fallback stage. Ignore
arbitrary provider-supplied stage properties and messages.
- Map only exact new codes to Rust diagnostic categories. Add real
ACPX-client and sidecar-pipe regressions with secret canaries and
unknown-code controls.
- Bind the changed shared source and new diagnostic modules in Copilot
profile 70. Preserve profile 69 and earlier history.
- Document the diagnostic contract. There are no dependency, lockfile,
workflow, auth, model, permission, or retry-policy changes.
## Verification
- Red before the production change: four TypeScript cases lost the
native reason or host stage. A Rust case returned `unclassified` for the
new closed categories.
- Final focused checks: 320 tests passed: 241 across sidecar
classification, Grok admission, the real ACPX client, recovery, runtime
adapter, and Copilot installation; all 79 host integration tests also
passed in isolation. These include nine startup-stage cases, translated
errors, actual missing Pi skill materialization, missing Copilot/Cursor
credentials, and a missing workspace binding.
- Review follow-up: a translated Copilot rejection lost its stage, and a
real Pi skill-loading failure reported configuration. Both regressions
failed before the two-line repair and pass afterward. Stage recording
now follows provider translation and starts before Pi skill acquisition.
The current unpublished Copilot 70 identity was regenerated.
- Formal review follow-up: missing Copilot and Cursor credentials
reported configuration. Both cases failed before the two-line stage
repair and pass afterward. A missing-workspace case also verifies the
separate binding catch.
- Seven Rust unit checks and 17 sidecar transport integration checks
passed. The pipe checks preserve known categories and reject unknown
suffixes without disclosing the provider message.
- `pnpm -r typecheck` and `pnpm build` passed. Rust formatting and the
generated profile/source identity checks passed.
- Host integration history is retained: one complete 73-case host run
passed. Other batches hit credential-port ownership errors in different
existing cases; the four new stage regressions passed. The final 79-case
host run passed after the last review fix. The prior intermittent lease
failures remain recorded and are not claimed fixed here.
- `pnpm test:run` was attempted and stopped after repeated embedded
Postgres bootstrap failures. The same local database failure predates
this change. Full CI must pass before merge.
- A metadata-only probe with pinned Grok 1.0.13, private HOME, synthetic
credentials, a local catalog, and no inference passed fresh Grok
4.7/High admission and cold restore. This is supporting ABI evidence. It
does not prove live subscription execution.
- Initial CI history is retained: server shard 1 failed its unchanged
final rate-limit slot test (two requests allowed instead of one). No
server policy code changed here. The review follow-up requires fresh
full CI; this is not claimed fixed.
- Second CI history is retained: chat shard 2 failed the unchanged Slack
session-sync quiet-queue test (three selects during the quiet window).
The first rate-limit failure did not recur there. No chat or server
policy code changed in this PR. Final head
654bc5aa21 passed all 61 contexts (56
success, four intentional skips, one informational neutral review). Both
final-head reviews are 5/5 with no new findings; the Paperclip app
publishes its completed review as an informational neutral check.
- Live native acceptance is still pending. After deployment, retry the
original bounded task and inspect its actual shell output, exit status,
run result, and saved task state. This PR fixes the evidence gap; it
does not claim the startup incident is resolved.
## Risks
- Shared startup failures can now receive a stage category. The private
weak map cannot import provider-defined fields. Known error codes keep
priority.
- A category could point at the wrong stage. Tests inject failures at
the installation, sandbox, credential, and runtime boundaries. Existing
cancellation and cleanup tests remain in place.
- The Copilot profile identity changes because shared source changes.
Historical identities remain decodable and cannot authorize a current
launch.
- The next live failure may require another functional repair. Admission
remains closed until the requested configuration is verified.
## Model Used
- OpenAI Codex, based on GPT-6, with reasoning, tool use, and code
execution. The exact deployment model ID and context-window size are not
exposed in this session.
## 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 (focused checks; full-suite
environment failures are recorded above)
- [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
- [x] All Paperclip CI gates are green
- [x] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge
---------
Co-authored-by: Paperclip <noreply@paperclip.ing>
11 KiB
Grok Build native runner
Select Grok Build in the native runner provider selector. The stored contract is
adapterType: "paperclip_runner" with provider: "acpx", acpxAgent: "grok",
and model: "grok-4.7". Existing grok_local agents keep their legacy adapter
until an explicit upgrade. On a qualified macOS ARM64 or Linux x64 computer,
Harness / Runtime → Upgrade runner preserves the agent, model, account,
instructions, and history. It starts fresh execution sessions. Advanced also
offers an explicit Runner choice. Unsupported legacy settings are rejected
before saving; select compatible settings or keep Legacy.
New Grok runner agents default to Full auto (approve all)
(acpxPermissionMode: "approve-all") in setup and the configuration form.
API configurations that omit the permission mode use the same default. No
additional permission setting is needed for unattended execution. Explicitly
saved restrictions remain unchanged.
Grok Build's qualified harness does not require the enableNativeRunner flag.
Explicit thinking effort uses reasoningEffort (minimal, low, medium,
high, or xhigh). The immutable v10 execution input, provider descriptor,
and recovery identity bind this value as grokEffort. The Grok provider boundary
adds native _meta.modelId and _meta.reasoningEffort to session create, load,
and resume requests before transport writes and observation. Grok 1.0.13 reports
the selected effort in _meta["x.ai/sessionConfig"].options; it does not support
the standard session/set_config_option effort API. The Runner checks that native
acknowledgement and the effective model before a prompt. A lazy restored handle
must connect through read-only native session inspection first. Reapplying the
bound metadata on restore prevents Grok from resetting effort to its default.
Missing configuration, a rejected value, or effort drift stops admission.
Admission failures keep a closed Grok reason code through the sidecar boundary.
Other startup failures retain their closed admission stage when no more specific
known error code is available. These diagnostics contain no native payload,
credential, arbitrary provider message, or workspace path. They do not change
admission or retry policy.
Omitting effort retains Grok's provider default. This setting remains outside
the model-visible task prompt. Historical execution inputs retain their schemas.
The upgrade resolves an unset model or the legacy grok-build default to the
native grok-4.7 default. Task-level model and effort overrides use the Grok
catalog and keep the account and tool policy unchanged.
Migration also preserves the effective disableWebSearch setting, whose legacy
default is true. Native input v10 carries it as grokDisableWebSearch through
startup and recovery. Grok's private requirements.toml enforces disabled search
above compatible project settings. Explicit false retains enabled search;
unset native configurations retain their provider default. See Grok's
configuration reference.
Grok Build speaks ACP over stdio.
The runner owns grok agent --no-leader stdio through ACPX, including session
identity, cancellation, recovery and the authenticated Paperclip MCP bridge.
ACP permission requests are approved by the runner under the default full-auto
policy. It does not add --always-approve: permission decisions remain under
the selected ACPX policy. Grok's ACP metadata cannot independently establish
Paperclip tool authority, so explicitly selecting approve-paperclip or
approve-reads returns the approval-required outcome; deny-all rejects requests.
Full auto does not bypass Paperclip's company permissions, governed approvals,
or execution-environment boundaries. Isolated ask
rules override project allow rules, and compatible always-approve settings are
locked off. Compatible hook/MCP discovery and shell login capture are disabled.
Installation and identity
Grok support ships inside the native runner and the public Paperclip server npm
artifact. There is no separate Grok npm package, binary payload, or npm lifecycle
download. The built-in launcher is identified as builtin:grok-acp version 1;
its native runtime identity is native:grok version 1.0.13. The historical
agentServerPackage/agentRuntimePackage wire fields carry these identities,
not npm dependencies. Existing npm-backed ACP bridges retain their package pins.
Provision Grok Build 1.0.13 at
/opt/paperclip/providers/grok/1.0.13/grok in the selected execution environment.
The sandbox provisioning helper is explicit and is never run by npm:
sudo node packages/paperclip-runner/scripts/provision-grok.mjs /opt/paperclip/providers/grok/1.0.13/grok
The standard Daytona image provisions it separately from the provider pack.
Custom images and local execution hosts must provide the same prerequisite.
The runner verifies the native executable checksum before credential refresh or
ACP startup; a missing prerequisite reports the required path and version.
Only macOS arm64 and Linux x64 are qualified. No ambient grok from PATH is used.
ACP must report the requested exact model; mismatches fail closed.
The distribution identity change deliberately rejects resume bindings from the former private-package profile. Start a fresh session after upgrading that unreleased profile; do not silently reinterpret its saved identity.
Instructions use Grok ACP session rules. Assigned skills live in the isolated Grok home. Steering and goals are unsupported. Token and cost values remain unknown when Grok does not report them; missing usage is not zero usage.
Authentication
Use the existing Grok company connection/login flow for subscription execution.
An explicitly selected company-secret XAI_API_KEY selects paid API execution.
There is no automatic subscription-to-API fallback. Remote execution cannot
borrow the operator's home credentials. Local execution can use the operator's
existing Grok login when no company login has been selected.
Only the selected credential is staged in the private runtime home. An ownership
lease fences concurrent processes. Before ACP startup, an expiring subscription
credential is refreshed through the verified Grok executable’s non-inference
models command. This bounded step suppresses output and prevents Grok 1.0.13
from caching a pre-refresh model list. Exact model verification still precedes
any prompt; refresh failure requires reconnecting Grok Build. After the provider exits, refreshed credentials
are copied back through Grok's existing identity and refresh checks. Runtime
credentials, refresh handoffs, and diagnostic logs are excluded from workspace
backups and removed on close. Session history remains available for resume. Host
configuration, other provider credentials and unselected keys are not forwarded
to Grok.
Evaluation
The private paperclip-evals repository maintains live-acpx-grok-4.7.json and
rosters/live-acpx-grok.json. The roster covers all 39 protocol cases. Its campaign
lane remains disabled pending complete live qualification.
Product E2E exposes runner-acpx-grok in core local/Daytona compatibility and
local session integrity. The explicit grok-qualification suite covers replies,
planning approvals, structured questions, downloadable project revisions, stop/resume
and continuation after controller restart
in both environments. Run with --suite grok-qualification; it is excluded from
scheduled --all. Use the canonical Product E2E dashboard and Evalbook reports.
The separate grok-subscription-qualification suite covers the same workflows
with the explicit GROK_AUTH_JSON fixture credential. It seeds only the disposable
company's private login home and supplies no API key. It does not exercise the
interactive sign-in UI. Authentication mode remains part of the profile identity.
After a subscription upgrade, a fresh grok login may be needed if the existing
login still reports the previous entitlement through ACP.
packages/paperclip-runner/scripts/grok-native-smoke.mjs records explicit auth,
model/binary identity, MCP outcome, durable resume and restrictive permissions.
Pass --auth subscription --auth-file /private/path/auth.json --output /private/report.json
or --auth api --output /private/report.json with an explicitly supplied key.
Each attempt is retained; unknown usage and cost are null.
Qualification requires the full protocol roster, selected product workflows, subscription and API execution locally and on Daytona, and three successful repetitions of core tool, approval and resume cases. Deterministic tests or a single successful browser task do not establish that qualification.
Remote verification
Do not run Docker on a developer laptop when using remote verification. The
Docker Runner check workflow offers maintainer-authorized manual Linux image
builds and broad source checks without provider credentials. It records the
source revision, resolved lock digest and immutable image reference. Paid
Product E2E remains behind the protected default-branch workflow and environment.
The public npm consumer check uses a digest-pinned, unprivileged container with
no checkout or credentials mounted. It downloads dependencies with lifecycle
scripts disabled and freezes the resulting consumer lockfile. It completes the
clean install with offline npm rebuild, running the deferred lifecycle hooks
without re-resolving bundled optional dependencies. Networking stays disabled,
the lockfile must remain unchanged, and a sentinel proves scripts actually ran.
The pinned image includes native build tools and local Node headers so dependency
hooks can compile without network access. Both executable admission probes run
in the same isolation. The verification user provisions Grok inside the disposable
test directory without privilege elevation or host /opt changes. Only the
positive probe mounts that binary read-only at the canonical sandbox path.
The Cloud application image also carries the controller-owned provider pack and
sets PAPERCLIP_RUNNER_REMOTE_PROVIDER_PACK_PATH. Remote ACPX execution verifies
the sandbox against that pack before using it, or stages the matching pack when
needed. The Cloud controller image does not install the native Grok executable;
the selected sandbox image must provide the prerequisite above.
Cloud builds must supply the full source SHA through PAPERCLIP_BUILD_COMMIT
to produce that verified pack. Unstamped local Cloud builds still work for other
features, but omit the pack and cannot start remote ACPX sessions.