Add low-trust review containment (#7530)

## Thinking Path

> - Paperclip is a control plane for AI-agent companies, so execution
policy and trust boundaries are part of the product's safety contract.
> - Low-trust review work needs narrower authority than normal
same-company agents because hostile PRs, comments, attachments, and
generated output can carry prompt-injection payloads.
> - The current V1 shape gives trusted workers broad company context,
which is useful for normal execution but too permissive for a reviewer
assigned to hostile content.
> - This branch adds a `low_trust_review` preset, source-trust tagging,
route-level containment, and quarantine handling so low-trust output
does not automatically flow into higher-trust wake context.
> - The branch has been rebased onto current `origin/master`, and the
low-trust migration was renumbered to `0097_low_trust_source_trust.sql`
to avoid collisions with existing `0091` through `0096` migrations.
> - Greptile feedback was addressed by tightening low-trust detection,
preserving project-level trust policy checks, fixing issue-kind
promotion lookup, removing duplicate post-lease isolation assertion,
documenting fail-closed source-trust behavior, bounding ancestry checks,
enforcing runtime issue context for CEOs, awaiting accepted-plan monitor
authorization, and making low-trust issue source-trust tagging atomic.
> - The benefit is a first production slice of deny-by-default review
containment with regression coverage for the main control-plane pivot
surfaces.

Fixes #7531.

## What Changed

- Added shared trust-policy types and validators, plus
database/source-trust fields for issues, comments, documents, and work
products.
- Implemented server enforcement for low-trust issue scope, agent
self-view redaction, secret/plugin/runtime denial paths, promotion
checks, and quarantined continuation/wake context.
- Added focused low-trust regression tests for resolver behavior, source
trust, route authorization, heartbeat preflight ordering, runtime
containment, and quarantine redaction.
- Added board UI affordances for selecting/reviewing the low-trust
preset and surfacing source-trust badges in relevant issue views.
- Added `doc/LOW-TRUST-PRESETS.md`, updated
`doc/SPEC-implementation.md`, and committed the low-trust review
contract plan under `doc/plans/`.
- Rebasing note: the original `0097_low_trust_source_trust.sql`
migration was renamed to `0097_low_trust_source_trust.sql`; the SQL uses
`ADD COLUMN IF NOT EXISTS` so users who already applied the old-numbered
migration are not broken by the renumbered migration.

## Verification

- Rebased branch onto current `origin/master` and force-pushed with
lease to `origin/PAP-10211-low-trust-agent` at head `2719f31e3`.
- Confirmed the PR diff does not include `pnpm-lock.yaml` or
`.github/workflows` changes.
- Resolved upstream UI/comment conflicts by preserving deleted-comment
tombstone behavior and low-trust source-trust badges/metadata.
- Renumbered the low-trust source-trust migration to
`0097_low_trust_source_trust.sql`; the SQL uses `ADD COLUMN IF NOT
EXISTS` so users who already applied an old-numbered copy are not
broken.
- `pnpm exec vitest run ui/src/lib/issue-chat-messages.test.ts
server/src/__tests__/heartbeat-workspace-session.test.ts`
- `pnpm exec vitest run server/src/__tests__/source-trust.test.ts
server/src/__tests__/workspace-runtime-service-authz.test.ts
ui/src/lib/trust-policy-ui.test.ts
ui/src/components/TrustPresetSection.test.tsx`
- `pnpm run typecheck:build-gaps`
- `git diff --check`
- GitHub checks pass on head `2719f31e3`: build, typecheck/release
registry, general tests, serialized server suites, e2e, canary, verify,
policy/review, Socket, and Snyk.
- Greptile Review passes with Confidence Score 5/5 and zero unresolved
Greptile review threads.
- No design screenshots/images were added because the task explicitly
says not to add them unless they are specifically part of the work.

## Risks

- Medium risk: this touches shared trust-policy contracts, server
authorization paths, heartbeat context generation, migration metadata,
and UI preset controls.
- Low-trust containment is intentionally deny-by-default; legitimate
future review workflows may need explicit allowlisted exceptions.
- Plugin/runtime/security surfaces are broad, so regression tests cover
the current known routes but future integrations must route through the
same containment layer.
- The PR is ready for review; GitHub checks are green and Greptile is
5/5.

> 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, GPT-5 coding agent, tool-enabled shell and GitHub CLI
workflow.

## 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] UI changes are covered by focused tests; no screenshots were added
per task instruction not to add design images unless specifically
required
- [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

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
This commit is contained in:
Dotta
2026-06-05 11:48:02 -10:00
committed by GitHub
parent b2a33d0184
commit dbebf30c89
87 changed files with 6522 additions and 121 deletions
@@ -0,0 +1,4 @@
ALTER TABLE "issues" ADD COLUMN IF NOT EXISTS "source_trust" jsonb;--> statement-breakpoint
ALTER TABLE "issue_comments" ADD COLUMN IF NOT EXISTS "source_trust" jsonb;--> statement-breakpoint
ALTER TABLE "documents" ADD COLUMN IF NOT EXISTS "source_trust" jsonb;--> statement-breakpoint
ALTER TABLE "issue_work_products" ADD COLUMN IF NOT EXISTS "source_trust" jsonb;
@@ -680,6 +680,13 @@
"when": 1780507597455,
"tag": "0096_document_annotation_issue_comment_links",
"breakpoints": true
},
{
"idx": 97,
"version": "7",
"when": 1780534000000,
"tag": "0097_low_trust_source_trust",
"breakpoints": true
}
]
}
+3 -1
View File
@@ -1,4 +1,5 @@
import { pgTable, uuid, text, integer, timestamp, index } from "drizzle-orm/pg-core";
import { pgTable, uuid, text, integer, timestamp, index, jsonb } from "drizzle-orm/pg-core";
import type { SourceTrustMetadata } from "@paperclipai/shared";
import { companies } from "./companies.js";
import { agents } from "./agents.js";
@@ -19,6 +20,7 @@ export const documents = pgTable(
lockedAt: timestamp("locked_at", { withTimezone: true }),
lockedByAgentId: uuid("locked_by_agent_id").references(() => agents.id, { onDelete: "set null" }),
lockedByUserId: text("locked_by_user_id"),
sourceTrust: jsonb("source_trust").$type<SourceTrustMetadata | null>(),
createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
updatedAt: timestamp("updated_at", { withTimezone: true }).notNull().defaultNow(),
},
+7 -1
View File
@@ -1,4 +1,9 @@
import type { IssueCommentAuthorType, IssueCommentMetadata, IssueCommentPresentation } from "@paperclipai/shared";
import type {
IssueCommentAuthorType,
IssueCommentMetadata,
IssueCommentPresentation,
SourceTrustMetadata,
} from "@paperclipai/shared";
import { pgTable, uuid, text, timestamp, index, jsonb } from "drizzle-orm/pg-core";
import { companies } from "./companies.js";
import { issues } from "./issues.js";
@@ -23,6 +28,7 @@ export const issueComments = pgTable(
deletedByAgentId: uuid("deleted_by_agent_id").references(() => agents.id, { onDelete: "set null" }),
deletedByUserId: text("deleted_by_user_id"),
deletedByRunId: uuid("deleted_by_run_id").references(() => heartbeatRuns.id, { onDelete: "set null" }),
sourceTrust: jsonb("source_trust").$type<SourceTrustMetadata | null>(),
createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
updatedAt: timestamp("updated_at", { withTimezone: true }).notNull().defaultNow(),
},
@@ -7,6 +7,7 @@ import {
timestamp,
uuid,
} from "drizzle-orm/pg-core";
import type { SourceTrustMetadata } from "@paperclipai/shared";
import { companies } from "./companies.js";
import { executionWorkspaces } from "./execution_workspaces.js";
import { heartbeatRuns } from "./heartbeat_runs.js";
@@ -36,6 +37,7 @@ export const issueWorkProducts = pgTable(
healthStatus: text("health_status").notNull().default("unknown"),
summary: text("summary"),
metadata: jsonb("metadata").$type<Record<string, unknown>>(),
sourceTrust: jsonb("source_trust").$type<SourceTrustMetadata | null>(),
createdByRunId: uuid("created_by_run_id").references(() => heartbeatRuns.id, { onDelete: "set null" }),
createdAt: timestamp("created_at", { withTimezone: true }).notNull().defaultNow(),
updatedAt: timestamp("updated_at", { withTimezone: true }).notNull().defaultNow(),
+2
View File
@@ -17,6 +17,7 @@ import { companies } from "./companies.js";
import { heartbeatRuns } from "./heartbeat_runs.js";
import { projectWorkspaces } from "./project_workspaces.js";
import { executionWorkspaces } from "./execution_workspaces.js";
import type { SourceTrustMetadata } from "@paperclipai/shared";
export const issues = pgTable(
"issues",
@@ -61,6 +62,7 @@ export const issues = pgTable(
.references((): AnyPgColumn => executionWorkspaces.id, { onDelete: "set null" }),
executionWorkspacePreference: text("execution_workspace_preference"),
executionWorkspaceSettings: jsonb("execution_workspace_settings").$type<Record<string, unknown>>(),
sourceTrust: jsonb("source_trust").$type<SourceTrustMetadata | null>(),
startedAt: timestamp("started_at", { withTimezone: true }),
completedAt: timestamp("completed_at", { withTimezone: true }),
cancelledAt: timestamp("cancelled_at", { withTimezone: true }),
+25
View File
@@ -7,6 +7,24 @@ export {
parseFrontmatterMarkdown,
type MarkdownDoc,
} from "./frontmatter.js";
export {
TRUST_PRESETS,
DEFAULT_TRUST_PRESET,
LOW_TRUST_REVIEW_PRESET,
LOW_TRUST_REVIEW_PRESET_VERSION,
LOW_TRUST_REVIEW_RAW_OUTPUT_DISPOSITION,
LOW_TRUST_TOOL_CLASSES,
type TrustPreset,
type LowTrustToolClass,
type LowTrustOutputPromotionTarget,
type LowTrustBoundary,
type LowTrustReviewPresetPolicy,
type TrustAuthorizationPolicy,
type SourceTrustArtifactKind,
type SourceTrustDisposition,
type SourceTrustPromotionSource,
type SourceTrustMetadata,
} from "./trust-policy.js";
export {
COMPANY_STATUSES,
DEFAULT_COMPANY_ATTACHMENT_MAX_BYTES,
@@ -817,8 +835,15 @@ export {
instanceExperimentalSettingsSchema,
patchInstanceExperimentalSettingsSchema,
issueGraphLivenessAutoRecoveryRequestSchema,
trustPresetSchema,
lowTrustBoundarySchema,
lowTrustReviewPresetPolicySchema,
trustAuthorizationPolicySchema,
type PatchInstanceExperimentalSettings,
type IssueGraphLivenessAutoRecoveryRequest,
type TrustPresetInput,
type LowTrustBoundaryInput,
type TrustAuthorizationPolicyInput,
} from "./validators/index.js";
export {
+67
View File
@@ -0,0 +1,67 @@
export const TRUST_PRESETS = ["standard", "low_trust_review"] as const;
export type TrustPreset = (typeof TRUST_PRESETS)[number];
export const DEFAULT_TRUST_PRESET = "standard" as const;
export const LOW_TRUST_REVIEW_PRESET = "low_trust_review" as const;
export const LOW_TRUST_REVIEW_PRESET_VERSION = 1 as const;
export const LOW_TRUST_REVIEW_RAW_OUTPUT_DISPOSITION = "quarantine" as const;
export const LOW_TRUST_TOOL_CLASSES = [
"git.read",
"github.pr.read",
"tests.local",
] as const;
export type LowTrustToolClass = (typeof LOW_TRUST_TOOL_CLASSES)[number] | string;
export interface LowTrustOutputPromotionTarget {
type: "issue";
issueId: string;
}
export interface LowTrustBoundary {
mode: typeof LOW_TRUST_REVIEW_PRESET;
companyId?: string;
projectIds?: string[];
rootIssueId?: string;
issueIds?: string[];
allowedAgentIds?: string[];
allowedSecretBindingIds?: string[];
allowedToolClasses?: LowTrustToolClass[];
outputPromotionTarget?: LowTrustOutputPromotionTarget;
}
export interface LowTrustReviewPresetPolicy {
id: typeof LOW_TRUST_REVIEW_PRESET;
version: typeof LOW_TRUST_REVIEW_PRESET_VERSION;
rawOutputDisposition: typeof LOW_TRUST_REVIEW_RAW_OUTPUT_DISPOSITION;
}
export interface TrustAuthorizationPolicy extends Record<string, unknown> {
trustPreset?: TrustPreset;
reviewPreset?: LowTrustReviewPresetPolicy;
trustBoundary?: LowTrustBoundary;
}
export type SourceTrustArtifactKind = "issue" | "comment" | "document" | "work_product";
export type SourceTrustDisposition = "quarantined" | "promoted";
export interface SourceTrustPromotionSource {
artifactKind: SourceTrustArtifactKind;
artifactId: string;
issueId?: string | null;
}
export interface SourceTrustMetadata {
preset: TrustPreset;
disposition: SourceTrustDisposition;
sourceIssueId?: string | null;
sourceRunId?: string | null;
sourceAgentId?: string | null;
promotedFrom?: SourceTrustPromotionSource | null;
promotedByActorType?: "agent" | "user" | "system" | null;
promotedByActorId?: string | null;
promotedAt?: string | null;
}
+7 -1
View File
@@ -9,9 +9,15 @@ import type {
CompanyMembership,
PrincipalPermissionGrant,
} from "./access.js";
import type {
TrustAuthorizationPolicy,
TrustPreset,
} from "../trust-policy.js";
export interface AgentPermissions {
export interface AgentPermissions extends Record<string, unknown> {
canCreateAgents: boolean;
trustPreset?: TrustPreset;
authorizationPolicy?: TrustAuthorizationPolicy;
}
export interface AgentModelProfileConfig {
+14
View File
@@ -40,6 +40,20 @@ export {
MIN_ISSUE_GRAPH_LIVENESS_AUTO_RECOVERY_LOOKBACK_HOURS,
MAX_ISSUE_GRAPH_LIVENESS_AUTO_RECOVERY_LOOKBACK_HOURS,
} from "./instance.js";
export {
TRUST_PRESETS,
DEFAULT_TRUST_PRESET,
LOW_TRUST_REVIEW_PRESET,
LOW_TRUST_REVIEW_PRESET_VERSION,
LOW_TRUST_REVIEW_RAW_OUTPUT_DISPOSITION,
LOW_TRUST_TOOL_CLASSES,
type TrustPreset,
type LowTrustToolClass,
type LowTrustOutputPromotionTarget,
type LowTrustBoundary,
type LowTrustReviewPresetPolicy,
type TrustAuthorizationPolicy,
} from "../trust-policy.js";
export type {
CompanySkillSourceType,
CompanySkillTrustLevel,
+10
View File
@@ -30,6 +30,11 @@ import type { Goal } from "./goal.js";
import type { Project, ProjectWorkspace } from "./project.js";
import type { ExecutionWorkspace, IssueExecutionWorkspaceSettings } from "./workspace-runtime.js";
import type { IssueWorkProduct } from "./work-product.js";
import type {
LowTrustReviewPresetPolicy,
SourceTrustMetadata,
TrustAuthorizationPolicy,
} from "../trust-policy.js";
export type { IssueWorkMode };
@@ -99,6 +104,7 @@ export interface IssueDocumentSummary {
lockedAt: Date | null;
lockedByAgentId: string | null;
lockedByUserId: string | null;
sourceTrust?: SourceTrustMetadata | null;
createdAt: Date;
updatedAt: Date;
}
@@ -441,6 +447,8 @@ export interface IssueExecutionPolicy {
commentRequired: boolean;
stages: IssueExecutionStage[];
monitor?: IssueExecutionMonitorPolicy | null;
reviewPreset?: LowTrustReviewPresetPolicy;
authorizationPolicy?: TrustAuthorizationPolicy;
}
export interface IssueExecutionMonitorState {
@@ -537,6 +545,7 @@ export interface Issue {
completedAt: Date | null;
cancelledAt: Date | null;
hiddenAt: Date | null;
sourceTrust?: SourceTrustMetadata | null;
labelIds?: string[];
labels?: IssueLabel[];
blockedBy?: IssueRelationIssueSummary[];
@@ -584,6 +593,7 @@ export interface IssueComment {
deletedByAgentId?: string | null;
deletedByUserId?: string | null;
deletedByRunId?: string | null;
sourceTrust?: SourceTrustMetadata | null;
followUpRequested?: boolean;
createdAt: Date;
updatedAt: Date;
@@ -49,6 +49,7 @@ export interface IssueWorkProduct {
healthStatus: "unknown" | "healthy" | "unhealthy";
summary: string | null;
metadata: Record<string, unknown> | null;
sourceTrust?: import("../trust-policy.js").SourceTrustMetadata | null;
createdByRunId: string | null;
createdAt: Date;
updatedAt: Date;
@@ -1,3 +1,5 @@
import type { TrustAuthorizationPolicy } from "../trust-policy.js";
export type ExecutionWorkspaceStrategyType =
| "project_primary"
| "git_worktree"
@@ -155,6 +157,7 @@ export interface ProjectExecutionWorkspacePolicy {
pullRequestPolicy?: Record<string, unknown> | null;
runtimePolicy?: Record<string, unknown> | null;
cleanupPolicy?: Record<string, unknown> | null;
authorizationPolicy?: TrustAuthorizationPolicy | null;
}
export interface IssueExecutionWorkspaceSettings {
+6 -1
View File
@@ -7,10 +7,13 @@ import {
} from "../constants.js";
import { agentAdapterTypeSchema } from "../adapter-type.js";
import { envConfigSchema } from "./secret.js";
import { trustAuthorizationPolicySchema, trustPresetSchema } from "./trust-policy.js";
export const agentPermissionsSchema = z.object({
canCreateAgents: z.boolean().optional().default(false),
});
trustPreset: trustPresetSchema.optional(),
authorizationPolicy: trustAuthorizationPolicySchema.optional(),
}).catchall(z.unknown());
export const agentInstructionsBundleModeSchema = z.enum(["managed", "external"]);
@@ -158,6 +161,8 @@ export type TestAdapterEnvironment = z.infer<typeof testAdapterEnvironmentSchema
export const updateAgentPermissionsSchema = z.object({
canCreateAgents: z.boolean(),
canAssignTasks: z.boolean(),
trustPreset: trustPresetSchema.optional(),
authorizationPolicy: trustAuthorizationPolicySchema.optional(),
});
export type UpdateAgentPermissions = z.infer<typeof updateAgentPermissionsSchema>;
+13
View File
@@ -146,6 +146,19 @@ export {
type CompanyPortabilityImport,
} from "./company-portability.js";
export {
trustPresetSchema,
lowTrustBoundarySchema,
lowTrustReviewPresetPolicySchema,
trustAuthorizationPolicySchema,
sourceTrustArtifactKindSchema,
sourceTrustMetadataSchema,
type TrustPresetInput,
type LowTrustBoundaryInput,
type TrustAuthorizationPolicyInput,
type SourceTrustMetadataInput,
} from "./trust-policy.js";
export {
createAgentSchema,
createAgentHireSchema,
+3
View File
@@ -27,6 +27,7 @@ import {
MODEL_PROFILE_KEYS,
} from "../constants.js";
import { multilineTextSchema } from "./text.js";
import { lowTrustReviewPresetPolicySchema, trustAuthorizationPolicySchema } from "./trust-policy.js";
export const issueBlockedInboxStateSchema = z.enum([
"needs_attention",
@@ -197,6 +198,8 @@ export const issueExecutionPolicySchema = z.object({
commentRequired: z.boolean().optional().default(true),
stages: z.array(issueExecutionStageSchema).default([]),
monitor: issueExecutionMonitorPolicySchema.optional().nullable(),
reviewPreset: lowTrustReviewPresetPolicySchema.optional(),
authorizationPolicy: trustAuthorizationPolicySchema.optional(),
});
export const issueExecutionMonitorStateSchema = z.object({
@@ -1,6 +1,7 @@
import { z } from "zod";
import { PROJECT_STATUSES } from "../constants.js";
import { envConfigSchema } from "./secret.js";
import { trustAuthorizationPolicySchema } from "./trust-policy.js";
const executionWorkspaceStrategySchema = z
.object({
@@ -26,6 +27,7 @@ export const projectExecutionWorkspacePolicySchema = z
pullRequestPolicy: z.record(z.string(), z.unknown()).optional().nullable(),
runtimePolicy: z.record(z.string(), z.unknown()).optional().nullable(),
cleanupPolicy: z.record(z.string(), z.unknown()).optional().nullable(),
authorizationPolicy: trustAuthorizationPolicySchema.optional().nullable(),
})
.strict();
@@ -0,0 +1,61 @@
import { z } from "zod";
import {
LOW_TRUST_REVIEW_PRESET,
LOW_TRUST_REVIEW_PRESET_VERSION,
LOW_TRUST_REVIEW_RAW_OUTPUT_DISPOSITION,
TRUST_PRESETS,
} from "../trust-policy.js";
export const trustPresetSchema = z.enum(TRUST_PRESETS);
export const lowTrustOutputPromotionTargetSchema = z.object({
type: z.literal("issue"),
issueId: z.string().uuid(),
}).strict();
export const lowTrustBoundarySchema = z.object({
mode: z.literal(LOW_TRUST_REVIEW_PRESET),
companyId: z.string().uuid().optional(),
projectIds: z.array(z.string().uuid()).optional(),
rootIssueId: z.string().uuid().optional(),
issueIds: z.array(z.string().uuid()).optional(),
allowedAgentIds: z.array(z.string().uuid()).optional(),
allowedSecretBindingIds: z.array(z.string().uuid()).optional(),
allowedToolClasses: z.array(z.string().trim().min(1)).optional(),
outputPromotionTarget: lowTrustOutputPromotionTargetSchema.optional(),
}).strict();
export const lowTrustReviewPresetPolicySchema = z.object({
id: z.literal(LOW_TRUST_REVIEW_PRESET),
version: z.literal(LOW_TRUST_REVIEW_PRESET_VERSION),
rawOutputDisposition: z.literal(LOW_TRUST_REVIEW_RAW_OUTPUT_DISPOSITION),
}).strict();
export const trustAuthorizationPolicySchema = z.object({
trustPreset: trustPresetSchema.optional(),
reviewPreset: lowTrustReviewPresetPolicySchema.optional(),
trustBoundary: lowTrustBoundarySchema.optional(),
}).catchall(z.unknown());
export const sourceTrustArtifactKindSchema = z.enum(["issue", "comment", "document", "work_product"]);
export const sourceTrustMetadataSchema = z.object({
preset: trustPresetSchema,
disposition: z.enum(["quarantined", "promoted"]),
sourceIssueId: z.string().uuid().nullable().optional(),
sourceRunId: z.string().uuid().nullable().optional(),
sourceAgentId: z.string().uuid().nullable().optional(),
promotedFrom: z.object({
artifactKind: sourceTrustArtifactKindSchema,
artifactId: z.string().uuid(),
issueId: z.string().uuid().nullable().optional(),
}).strict().nullable().optional(),
promotedByActorType: z.enum(["agent", "user", "system"]).nullable().optional(),
promotedByActorId: z.string().trim().min(1).nullable().optional(),
promotedAt: z.string().datetime({ offset: true }).nullable().optional(),
}).strict();
export type TrustPresetInput = z.infer<typeof trustPresetSchema>;
export type LowTrustBoundaryInput = z.infer<typeof lowTrustBoundarySchema>;
export type TrustAuthorizationPolicyInput = z.infer<typeof trustAuthorizationPolicySchema>;
export type SourceTrustMetadataInput = z.infer<typeof sourceTrustMetadataSchema>;