Files
PaperClipAI/maps/agent-access-map/instance-operations.md
T
DottaandPaperclip 61bd4ac691 feat(agents): set up Antigravity with saved Google subscriptions (#18189)
## Thinking Path

> - Paperclip manages agents and their work.
> - Users choose providers and managed accounts while adding agents.
> - Antigravity was missing from the ordinary additional-agent picker
and its saved Google path did not complete native readiness.
> - A user needs a successful Test, an enabled task action, and a stable
account binding.
> - This pull request connects the ordinary setup flow to the native
candidate.
> - It also preserves retired credential refreshes across database
failures and server restarts.

## Linked Issues or Issue Description

Built on merged #18188 and #18187. Refs: #12503, #14938 for related
legacy CLI proposals.

**Agent or provider**

Google Antigravity ACP 1.4.3.

**Why this adapter is useful**

Users need to set up an Antigravity agent and run real work through
their managed Google subscription.

**How the agent is invoked**

Native Paperclip Runner uses the pinned Antigravity ACP runtime. Exact
operator admission remains required; the local candidate result does not
enable general release.

## What Changed

- Show Google Antigravity in the ordinary adapter grid and preserve the
draft choice when setup resumes.
- Reuse the same user’s saved personal default Google subscription and
discover the exact low model through the native account.
- Run a prompt-free native Test and wait for authoritative readiness
before enabling task assignment.
- Carry the managed account through native runs and recover retired
refreshes through the existing encrypted, fsynced journal with exact
generation checks.
- Retain the latest rotated credential through verification, storage,
and partial home-removal failures. Retry the exact account’s retained
cleanup before another provider use.
- Reconcile cleanup before taking the credential mutation lock, fence
owners arriving during lock acquisition, and support nested lock holders
without a join cycle.
- Bound sign-in retirement outside database transactions and preserve
claims for the existing cleanup owner; one pending sign-in does not
block unrelated accounts.
- Cover Gemini and Pi API-key recovery, keep API-key bulk manual-model
editing usable, and keep adapter discovery available after candidate
expiry.
- Skip taskless native timer work before run or credential admission.
Preserve current master monitor and deadline scheduling.

## Verification

- The registered local saved-Google-subscription journey passed all 16
matchers in 196,520 ms using eval source `12aba` against actual serving
source `4a37ee7c`. Both turns stayed in one conversation, produced exact
downloaded files of 61 and 119 bytes after reload, and used distinct
native primary processes. All observed owners retired, leases released,
and credential fences settled. Independent readback confirmed Done work,
no active owned runs, a connected saved account, and the retained agent
paused for inspection.
- The measured runtime is production-equivalent to pre-restack
`65ef31dc` except its test-only transport fixture correction.
Current-master head `ff7c9249` preserves all 99 product paths: 98 exact
blobs and only eight inherited plugin-panel lines in `agents.ts`. Its
native binary was rebuilt for inherited Codex-only diagnostic changes;
the changed Rust behavior is outside Antigravity's ACPX execution path.
Shared native/UI changes have focused regression coverage. The recorded
live result is not represented as a measured replay of `ff7c9249`.
- Full workspace build and typecheck passed on `ff7c9249`. All 376
focused tests in 14 files passed, covering Google credential/retirement
controls and the inherited native-tool, redaction, route, and UI
recovery changes. Token gates, maps, and diff checks passed. The
required-close fixture was separately demonstrated red, then passed all
317 selected provider-wiring cases.
- [Current-head
CI](https://github.com/paperclipai/paperclip/actions/runs/40959387972)
is green: 55 successful checks, two skipped workflow checks, and an
optional neutral review check. Required `ci / verify` and `ci / e2e`
passed. Greptile and Paperclip Review both report 5/5 on exact
`ff7c9249`, with zero unresolved review threads.
- This qualifies the selected operator-authorized local
saved-subscription candidate. The original exploratory journey remains
historical evidence; the repeatable fixed-driver eval is in the next PR.

## Risks

- Google refresh recovery must preserve the exact grant, account, and
credential generation. An uncertain native owner or failed encrypted
handoff retains the private home and blocks cleanup.
- A successful local saved-subscription candidate does not qualify fresh
OAuth, API keys, Cloud, Linux, other models, or general release.
- Timer filtering changes only taskless native Runner admissions.
Existing task-bound monitors and legacy adapters retain their owners.

## Model Used

OpenAI Codex (GPT-6). The host did not expose the exact model revision
or context limit. The model used tools for code edits, tests, and
review.

## 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
- [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>
2026-10-29 16:51:50 -05:00

10 KiB
Raw Blame History

Installation, configuration, updates, and backups — agent access

Product map: Installation, configuration, updates, and backups.

Feature ID: instance-operations.

Review date: 2026-10-12. Source commit: f981fa3cc10cb8663a87dec60345fbff46d54d0a.

This is a source-reviewed baseline of representative operations and known gaps. Each capability needs operation-level review before its entire scope can be qualified. No live agent execution is claimed. Follow the map contract for format, permission checks, evidence and maintenance.

install-update

Capability ID: instance-operations/install-update.

Review: baseline.

API: not-applicable — No uniform company REST updater API. Treat pre-start installation as an operator exception or design an explicit delegated operator API.

Agent access: operator-only — Existing CLI install/update/rollback commands.

Policy: Local machine operator and instance-administrator authority; distinct from company-scoped agent keys.

Evidence: source-only — implementation at the reviewed commit above.

Verification: not-run — exercise the listed local workflow in an authorized, owned operator environment. Verify its isolation and result; do not count local machine authority as company-agent API access.

Gaps: No uniform company REST updater API. Treat pre-start installation as an operator exception or design an explicit delegated operator API.

server-service

Capability ID: instance-operations/server-service.

Review: baseline.

API: present — representative operation IDs:

  • GET /api/health.

Agent access: operator-only — Existing CLI service/dev controls; health REST; owned workspace services have separate agent APIs.

Policy: Local machine operator and instance-administrator authority; distinct from company-scoped agent keys.

Evidence: source-only — implementation at the reviewed commit above.

Verification: not-run — exercise the listed local workflow in an authorized, owned operator environment. Verify its isolation and result; do not count local machine authority as company-agent API access.

Gaps: Instance start/stop and task workspace start/stop are different capabilities. Local service control is not agent company authority.

configuration

Capability ID: instance-operations/configuration.

Review: baseline.

API: present — representative operation IDs:

  • GET /api/instance/settings.
  • PATCH /api/instance/settings.
  • GET /api/instance/settings/experimental.
  • PATCH /api/instance/settings/experimental.

Agent access: restricted — Board/admin instance settings; CLI configure/env; local configuration files.

Policy: Local machine operator and instance-administrator authority; distinct from company-scoped agent keys.

Evidence: source-only — implementation at the reviewed commit above.

Verification: not-run — assert the documented agent denial with an owned test fixture; verify any listed agent request/result path separately. A delegated operation needs its own policy and access review.

Gaps: Settings APIs cover parts of configuration; pre-start database/binding/storage configuration is local operator tooling.

Cloud stack Admins may update experimental settings through the trusted Cloud actor without receiving other instance-admin powers. This uses the current stack role on each request; company membership alone does not grant it. Members, Support users, agents, and ordinary self-hosted sessions remain denied. Cloud staging exposes experimental controls to instance admins and stack Admins. It permits saved overrides of Cloud-disabled experimental defaults. Cloud-enabled features remain true on every managed stack, including staging, and the settings service rejects attempts to disable them with 403 before saving. This exception requires a managed instance and a provisioner-supplied *.staging.paperclip.app primary hostname; request headers do not establish staging. Ordinary board members and agents still cannot mutate instance settings. Only an explicit feature allowlist is editable on staging. Cloud owner-admin elevation and managed-sandbox execution policy remain managed, and production policy and other Cloud floors remain. Local route tests cover stack Admin experiment access and auditing, denial for other instance controls and actor sources, hidden-setting restrictions, and a spoofed request hostname; the Cloud actor tests cover role changes and preserved instance-admin isolation; no live Cloud or agent-credential verification is claimed.

diagnostics

Capability ID: instance-operations/diagnostics.

Review: baseline.

API: present — representative operation IDs:

  • GET /api/health.
  • GET /api/instance/task-drain?upgradeReadiness=1 — instance-admin or signed Cloud control only. Returns aggregate restart eligibility, never checkpoint contents, task identities or credentials. Company members and agent keys are denied. The explicit request performs bounded read-only SQL; ordinary drain status remains an in-memory read. Route coverage and PostgreSQL coverage establish this local boundary; hosted active-run continuity requires separate proof.
  • GET /api/liveness and HEAD /api/liveness — public process startup status and build commit only; no database readiness or workspace identity.

Liveness is anonymous and runs before actor/session resolution. Hostname and signed-bootstrap endpoint validation still apply. It grants no instance control authority. Liveness tests cover idle drain, invalid ambient credentials, unavailable SQL, and startup recovery with local Express fixtures; no live provider or agent access is claimed.

Agent access: partial — REST health plus local CLI doctor.

Policy: Local machine operator and instance-administrator authority; distinct from company-scoped agent keys.

Evidence: source-only — implementation at the reviewed commit above.

Verification: not-run — exercise a listed permitted operation with actual agent credentials/run context; inspect its result and applicable persistence/attribution. Assert the relevant denied actor, owner, company or mode from Policy. Review the remaining operations before qualifying the whole capability.

Gaps: Health is an API; full local doctor/repair inspects host configuration and is not equivalent to that response.

backup

Capability ID: instance-operations/backup.

Review: baseline.

API: present — representative operation IDs:

  • POST /api/instance/database-backups.

Agent access: restricted — Board/admin backup API; local CLI db:backup; restore into an authorized separate target.

Policy: Local machine operator and instance-administrator authority; distinct from company-scoped agent keys.

Evidence: source-only — implementation at the reviewed commit above.

Verification: not-run — assert the documented agent denial with an owned test fixture; verify any listed agent request/result path separately. A delegated operation needs its own policy and access review.

Gaps: API creation exists. Restore/lifecycle are operator workflows rather than ordinary company agent capabilities.

runtime-setup

Capability ID: instance-operations/runtime-setup.

Review: baseline.

API: not-applicable — these commands provision an operator-owned published package and local filesystem cache. They have no company-domain REST equivalent and deliberately bypass instance config, environment, and credential initialization.

Agent access: operator-only — explicit host/package provisioning through the runtime CLI, with the installed release’s verified assets. A company agent grant does not authorize another host’s installation or private operator cache.

Policy: Host setup verifies public pinned assets without model calls and requires the published server package layout/identity. Import requires a trusted manifest digest, exact installed source/profile identity, Linux x64 companion inventory/pack/ELF integrity, and a private operator-owned cache. Import performs no downloads or executable probes. A Pi companion does not enable remote Cursor or qualify credentials.

Evidence: source-only — reviewed at 9ef434fe79c57d976b8b03b14c5242db270b8320. Runtime CLI, instance-config bypass, Pi provisioning, Cursor provisioning, and companion importer.

Verification: not-run — Inspect setup identity/path and repeat reuse on an owned disposable published installation. Verify a matching trusted companion import and reject altered/cross-release/linked inventory while preserving unrelated files. Check no instance credentials are loaded and no model is called. Qualify actual adapter/account task execution separately.

Gaps: Source-checkout setup, different platforms, remote Cursor, provider execution, and published-package acceptance remain separate scopes. Local fixture materialization does not establish remote model readiness.

Antigravity addition (2026-10-28): paperclipai runtime setup antigravity stages the verified official and patched two-executable closures on macOS ARM64. The operator-only authority above is unchanged. The native profile remains pending at API, driver, sidecar and Rust admission; a company grant cannot enable it. Google AI API-key binding is distinct from Workspace OAuth. Evidence remains source-only for agent access: setup, provisioner, and gate regressions. Credential-free public-layout provisioning and controlled native tests do not qualify a company agent task or authenticated connection/revocation.