Guard cheap recovery model usage (#6371)

## Thinking Path

> - Paperclip is the control plane that coordinates AI-agent work
through issues, heartbeats, comments, approvals, and auditable recovery
paths.
> - The affected subsystem is heartbeat/recovery orchestration,
especially the optional cheap model profile used for operational
recovery overhead.
> - Cheap recovery should repair status and liveness, but it must not
become the worker lane that writes deliverables, continues source work,
or propagates cheap execution hints into downstream retries.
> - The gap was that cheap-profile hints could follow recovery wake
contexts and assignment overrides farther than intended, making real
work eligible to run on the cheap model.
> - This pull request separates status-only cheap recovery from normal
source-work continuations, adds route guards for deliverable mutations
during cheap status-only runs, and documents the invariant.
> - The benefit is safer retry/recovery behavior: cheap runs can clean
up control-plane state, while any remaining source work resumes through
a normal/original model path.

## What Changed

- Added recovery model-profile work classes so status-only recovery
carries explicit guard context and normal-model continuations scrub
cheap hints.
- Updated heartbeat, productivity review, liveness continuation, and
recovery service wakeups to request cheap only for bounded status-only
recovery work.
- Blocked cheap status-only recovery runs from writing issue documents,
plans, attachments, work products, or assigning downstream work back to
`modelProfile: "cheap"`.
- Added/updated server tests for cheap profile propagation,
artifact/document guards, route authorization, retry scheduling, and
successful-run handoff behavior.
- Documented the recovery model-profile lane in
`doc/SPEC-implementation.md` and `doc/execution-semantics.md`.
- After rebasing onto current `public-gh/master`, stabilized the new
`InstanceSidebar` plugin-filter tests so the PR check lane stays green.

## Verification

- Local: `pnpm exec vitest run --config vitest.config.ts
src/services/recovery/model-profile-hint.test.ts
src/__tests__/issue-agent-mutation-ownership-routes.test.ts
src/__tests__/issue-document-restore-routes.test.ts` from `server/` - 3
files, 37 tests passed after final edits.
- Local: `pnpm exec vitest run --config vitest.config.ts
src/__tests__/heartbeat-process-recovery.test.ts` from `server/` - 44
tests passed after rerunning the cleanup-sensitive file alone.
- Local: `pnpm --filter @paperclipai/ui exec vitest run
src/components/InstanceSidebar.test.tsx` - 4 tests passed.
- Local: `pnpm --filter @paperclipai/server typecheck` - passed.
- Local: `pnpm --filter @paperclipai/ui typecheck` - passed.
- PR checks on latest head `6f8c3b1380f5bd872c6f49f6f7188ecf3bb6d263` -
all green, including `verify`, build, typecheck,
server/general/serialized tests, e2e, Snyk, and policy.
- Greptile: pass 3 returned Confidence Score 5/5 with zero unresolved
Greptile review threads.

## Risks

- Medium risk: recovery behavior is intentionally stricter, so any path
that incorrectly relies on cheap recovery to keep doing source work will
now need to hand back to a normal-model run.
- Low migration risk: no schema changes.
- No product UI changes; the UI file touched is a test-only
stabilization after rebasing onto current `master`.

> For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and
discuss it in `#dev` before opening the PR. Feature PRs that overlap
with planned core work may need to be redirected — check the roadmap
first. See `CONTRIBUTING.md`.

## Model Used

- OpenAI Codex coding agent, GPT-5 model family (`gpt-5`), tool use and
local code execution enabled; context window not exposed in this
environment.

## 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 run tests locally and they pass
- [x] I have added or updated tests where applicable
- [x] If this change affects the UI, I have included before/after
screenshots (N/A: no product UI changes)
- [x] I have updated relevant documentation to reflect my changes
- [x] I have considered and documented any risks above
- [x] I will address all Greptile and reviewer comments before
requesting merge
This commit is contained in:
Dotta 2026-05-19 13:46:02 -05:00 committed by GitHub
parent 24748de421
commit bfe6369ef5
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
17 changed files with 529 additions and 78 deletions

View file

@ -7,6 +7,7 @@ import type { Db } from "@paperclipai/db";
import {
activityLog,
executionWorkspaces,
heartbeatRuns,
issueExecutionDecisions,
issueRelations,
issues as issueRows,
@ -1331,6 +1332,87 @@ export function issueRoutes(
return true;
}
function isStatusOnlyCheapRecoveryContext(contextSnapshot: unknown) {
if (!contextSnapshot || typeof contextSnapshot !== "object" || Array.isArray(contextSnapshot)) return false;
const context = contextSnapshot as Record<string, unknown>;
return context.modelProfile === "cheap" &&
context.recoveryIntent === "status_only" &&
context.allowDeliverableWork === false &&
context.allowDocumentUpdates === false &&
context.resumeRequiresNormalModel === true;
}
function requestsCheapIssueAssigneeModelProfile(input: { assigneeAdapterOverrides?: unknown }) {
const overrides = input.assigneeAdapterOverrides;
return !!overrides &&
typeof overrides === "object" &&
!Array.isArray(overrides) &&
(overrides as Record<string, unknown>).modelProfile === "cheap";
}
async function loadActorRunContext(req: Request, companyId: string) {
if (req.actor.type !== "agent") return null;
const runId = req.actor.runId?.trim();
if (!runId) return null;
const run = await db
.select({
id: heartbeatRuns.id,
companyId: heartbeatRuns.companyId,
agentId: heartbeatRuns.agentId,
contextSnapshot: heartbeatRuns.contextSnapshot,
})
.from(heartbeatRuns)
.where(eq(heartbeatRuns.id, runId))
.then((rows) => rows[0] ?? null);
if (!run || run.companyId !== companyId || run.agentId !== req.actor.agentId) return null;
return run;
}
async function assertCheapRecoveryIssueAssigneeProfileAllowed(
req: Request,
res: Response,
issue: { id?: string; companyId: string },
input: { assigneeAdapterOverrides?: unknown },
) {
if (!requestsCheapIssueAssigneeModelProfile(input)) return true;
const run = await loadActorRunContext(req, issue.companyId);
if (!run || !isStatusOnlyCheapRecoveryContext(run.contextSnapshot)) return true;
res.status(403).json({
error: "Cheap status-only recovery runs cannot assign downstream issue work to the cheap model profile",
details: {
issueId: issue.id ?? null,
runId: run.id,
modelProfile: "cheap",
recoveryIntent: "status_only",
resumeRequiresNormalModel: true,
},
});
return false;
}
async function assertDeliverableMutationAllowedByRunContext(
req: Request,
res: Response,
issue: { id: string; companyId: string },
) {
const run = await loadActorRunContext(req, issue.companyId);
if (!run) return true;
if (!isStatusOnlyCheapRecoveryContext(run.contextSnapshot)) return true;
res.status(403).json({
error: "Cheap status-only recovery runs cannot update issue documents, plans, or deliverable artifacts",
details: {
issueId: issue.id,
runId: run.id,
modelProfile: "cheap",
recoveryIntent: "status_only",
resumeRequiresNormalModel: true,
},
});
return false;
}
function assertStructuredCommentFieldsAllowed(
req: Request,
res: Response,
@ -2319,6 +2401,7 @@ export function issueRoutes(
}
assertCompanyAccess(req, issue.companyId);
if (!(await assertAgentIssueMutationAllowed(req, res, issue))) return;
if (!(await assertDeliverableMutationAllowedByRunContext(req, res, issue))) return;
const keyParsed = issueDocumentKeySchema.safeParse(String(req.params.key ?? "").trim().toLowerCase());
if (!keyParsed.success) {
res.status(400).json({ error: "Invalid document key", details: keyParsed.error.issues });
@ -2523,6 +2606,7 @@ export function issueRoutes(
}
assertCompanyAccess(req, issue.companyId);
if (!(await assertAgentIssueMutationAllowed(req, res, issue))) return;
if (!(await assertDeliverableMutationAllowedByRunContext(req, res, issue))) return;
const keyParsed = issueDocumentKeySchema.safeParse(String(req.params.key ?? "").trim().toLowerCase());
if (!keyParsed.success) {
res.status(400).json({ error: "Invalid document key", details: keyParsed.error.issues });
@ -2682,6 +2766,7 @@ export function issueRoutes(
}
assertCompanyAccess(req, issue.companyId);
if (!(await assertAgentIssueMutationAllowed(req, res, issue))) return;
if (!(await assertDeliverableMutationAllowedByRunContext(req, res, issue))) return;
const product = await workProductsSvc.createForIssue(issue.id, issue.companyId, {
...req.body,
projectId: req.body.projectId ?? issue.projectId ?? null,
@ -2725,6 +2810,7 @@ export function issueRoutes(
return;
}
if (!(await assertAgentIssueMutationAllowed(req, res, issue))) return;
if (!(await assertDeliverableMutationAllowedByRunContext(req, res, issue))) return;
const product = await workProductsSvc.update(id, req.body);
if (!product) {
res.status(404).json({ error: "Work product not found" });
@ -2765,6 +2851,7 @@ export function issueRoutes(
return;
}
if (!(await assertAgentIssueMutationAllowed(req, res, issue))) return;
if (!(await assertDeliverableMutationAllowedByRunContext(req, res, issue))) return;
const removed = await workProductsSvc.remove(id);
if (!removed) {
res.status(404).json({ error: "Work product not found" });
@ -2998,6 +3085,7 @@ export function issueRoutes(
const companyId = req.params.companyId as string;
assertCompanyAccess(req, companyId);
assertNoAgentHostWorkspaceCommandMutation(req, collectIssueWorkspaceCommandPaths(req.body));
if (!(await assertCheapRecoveryIssueAssigneeProfileAllowed(req, res, { companyId }, req.body))) return;
if (req.body.assigneeAgentId || req.body.assigneeUserId) {
await assertCanAssignTasks(req, companyId);
}
@ -3093,6 +3181,7 @@ export function issueRoutes(
}
assertCompanyAccess(req, parent.companyId);
assertNoAgentHostWorkspaceCommandMutation(req, collectIssueWorkspaceCommandPaths(req.body));
if (!(await assertCheapRecoveryIssueAssigneeProfileAllowed(req, res, parent, req.body))) return;
if (req.body.assigneeAgentId || req.body.assigneeUserId) {
await assertCanAssignTasks(req, parent.companyId);
}
@ -3239,6 +3328,7 @@ export function issueRoutes(
assertCompanyAccess(req, existing.companyId);
assertNoAgentHostWorkspaceCommandMutation(req, collectIssueWorkspaceCommandPaths(req.body));
if (!(await assertAgentIssueMutationAllowed(req, res, existing))) return;
if (!(await assertCheapRecoveryIssueAssigneeProfileAllowed(req, res, existing, req.body))) return;
const actor = getActorInfo(req);
const isClosed = isClosedIssueStatus(existing.status);
@ -5261,6 +5351,7 @@ export function issueRoutes(
return;
}
if (!(await assertAgentIssueMutationAllowed(req, res, issue))) return;
if (!(await assertDeliverableMutationAllowedByRunContext(req, res, issue))) return;
const company = await companiesSvc.getById(companyId);
const attachmentMaxBytes = normalizeIssueAttachmentMaxBytes(company?.attachmentMaxBytes);
@ -5380,6 +5471,7 @@ export function issueRoutes(
return;
}
if (!(await assertAgentIssueMutationAllowed(req, res, issue))) return;
if (!(await assertDeliverableMutationAllowedByRunContext(req, res, issue))) return;
try {
await storage.deleteObject(attachment.companyId, attachment.objectKey);