Files
paperclip/server/src/services/trust-preset-resolver.ts
T
Dotta dbebf30c89 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>
2026-06-05 16:48:02 -05:00

350 lines
11 KiB
TypeScript

import {
DEFAULT_TRUST_PRESET,
LOW_TRUST_REVIEW_PRESET,
type LowTrustBoundary,
type TrustPreset,
lowTrustBoundarySchema,
lowTrustReviewPresetPolicySchema,
trustAuthorizationPolicySchema,
trustPresetSchema,
} from "@paperclipai/shared";
type JsonRecord = Record<string, unknown>;
export const LOW_TRUST_ISSUE_ANCESTRY_MAX_DEPTH = 12;
export type TrustPresetPolicySource = "agent" | "project" | "issue" | "run";
export type ResolveCoreTrustPresetInput = {
companyId: string;
agent?: {
companyId?: string | null;
permissions?: unknown;
} | null;
project?: {
companyId?: string | null;
executionWorkspacePolicy?: unknown;
} | null;
issue?: {
companyId?: string | null;
executionPolicy?: unknown;
} | null;
run?: {
companyId?: string | null;
executionPolicy?: unknown;
} | null;
};
export type TrustPresetDenyReason =
| "unsupported_trust_preset"
| "invalid_authorization_policy"
| "invalid_low_trust_boundary"
| "cross_company_boundary"
| "conflicting_low_trust_boundary"
| "missing_low_trust_boundary_scope";
export type TrustPresetResolution =
| {
kind: "standard";
preset: typeof DEFAULT_TRUST_PRESET;
boundary: null;
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>;
}
| {
kind: "low_trust_review";
preset: typeof LOW_TRUST_REVIEW_PRESET;
boundary: LowTrustBoundary & { companyId: string };
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>;
}
| {
kind: "denied";
reason: TrustPresetDenyReason;
source: TrustPresetPolicySource | null;
detail: string;
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>;
};
type ParsedPolicySource = {
source: TrustPresetPolicySource;
companyId: string | null;
rawPolicy: JsonRecord | null;
authorizationPolicy: JsonRecord | null;
trustPreset: TrustPreset | null;
boundary: LowTrustBoundary | null;
impliesLowTrust: boolean;
};
function asRecord(value: unknown): JsonRecord | null {
return typeof value === "object" && value !== null && !Array.isArray(value)
? value as JsonRecord
: null;
}
function deny(
reason: TrustPresetDenyReason,
source: TrustPresetPolicySource | null,
detail: string,
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>,
): TrustPresetResolution {
return { kind: "denied", reason, source, detail, sourcePresets };
}
function isTrustPresetResolution(value: unknown): value is TrustPresetResolution {
if (!value || typeof value !== "object") return false;
const kind = (value as { kind?: unknown }).kind;
return kind === "standard" || kind === "low_trust_review" || kind === "denied";
}
function parsePreset(
value: unknown,
source: TrustPresetPolicySource,
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>,
): TrustPresetResolution | TrustPreset | null {
if (value === undefined || value === null) return null;
const parsed = trustPresetSchema.safeParse(value);
if (!parsed.success) {
return deny(
"unsupported_trust_preset",
source,
`Unsupported trust preset in ${source} policy.`,
sourcePresets,
);
}
return parsed.data;
}
function parseReviewPresetId(
value: unknown,
source: TrustPresetPolicySource,
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>,
): TrustPresetResolution | TrustPreset | null {
if (value === undefined || value === null) return null;
const parsed = lowTrustReviewPresetPolicySchema.safeParse(value);
if (!parsed.success) {
return deny(
"unsupported_trust_preset",
source,
`Unsupported review preset in ${source} policy.`,
sourcePresets,
);
}
return parsed.data.id;
}
function parseAuthorizationPolicy(
value: unknown,
source: TrustPresetPolicySource,
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>,
): TrustPresetResolution | JsonRecord | null {
if (value === undefined || value === null) return null;
const parsed = trustAuthorizationPolicySchema.safeParse(value);
if (!parsed.success) {
return deny(
"invalid_authorization_policy",
source,
`Invalid authorization policy in ${source} policy.`,
sourcePresets,
);
}
return parsed.data as JsonRecord;
}
function parseBoundary(
value: unknown,
source: TrustPresetPolicySource,
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>,
): TrustPresetResolution | LowTrustBoundary | null {
if (value === undefined || value === null) return null;
const parsed = lowTrustBoundarySchema.safeParse(value);
if (!parsed.success) {
return deny(
"invalid_low_trust_boundary",
source,
`Invalid low-trust boundary in ${source} policy.`,
sourcePresets,
);
}
return parsed.data;
}
function parseSource(
source: TrustPresetPolicySource,
companyId: string | null | undefined,
rawPolicy: JsonRecord | null,
authorizationPolicyInput: unknown,
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>,
): ParsedPolicySource | TrustPresetResolution {
const topPreset = parsePreset(rawPolicy?.trustPreset, source, sourcePresets);
if (isTrustPresetResolution(topPreset)) return topPreset;
const topReviewPreset = parseReviewPresetId(rawPolicy?.reviewPreset, source, sourcePresets);
if (isTrustPresetResolution(topReviewPreset)) return topReviewPreset;
const authorizationPolicy = parseAuthorizationPolicy(authorizationPolicyInput, source, sourcePresets);
if (isTrustPresetResolution(authorizationPolicy)) return authorizationPolicy;
const authPreset = parsePreset(authorizationPolicy?.trustPreset, source, sourcePresets);
if (isTrustPresetResolution(authPreset)) return authPreset;
const authReviewPreset = parseReviewPresetId(authorizationPolicy?.reviewPreset, source, sourcePresets);
if (isTrustPresetResolution(authReviewPreset)) return authReviewPreset;
const boundary = parseBoundary(authorizationPolicy?.trustBoundary, source, sourcePresets);
if (isTrustPresetResolution(boundary)) return boundary;
const trustPreset = topPreset ?? topReviewPreset ?? authPreset ?? authReviewPreset;
if (trustPreset) sourcePresets[source] = trustPreset;
return {
source,
companyId: companyId ?? null,
rawPolicy,
authorizationPolicy,
trustPreset,
boundary,
impliesLowTrust: trustPreset === LOW_TRUST_REVIEW_PRESET || Boolean(boundary),
};
}
function normalizeSet(values: readonly string[] | undefined): string[] | undefined {
if (values === undefined) return undefined;
return [...new Set(values)].sort();
}
function intersectSets(left: string[] | undefined, right: string[] | undefined): string[] | undefined {
const normalizedRight = normalizeSet(right);
if (normalizedRight === undefined) return left;
const normalizedLeft = normalizeSet(left);
if (normalizedLeft === undefined) return normalizedRight;
const rightSet = new Set(normalizedRight);
return normalizedLeft.filter((value) => rightSet.has(value));
}
function mergeBoundary(
current: (LowTrustBoundary & { companyId: string }) | null,
next: LowTrustBoundary,
companyId: string,
source: TrustPresetPolicySource,
sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>>,
): (LowTrustBoundary & { companyId: string }) | TrustPresetResolution {
if (next.companyId && next.companyId !== companyId) {
return deny(
"cross_company_boundary",
source,
"Low-trust boundary refers to a different company.",
sourcePresets,
);
}
const base = current ?? { mode: LOW_TRUST_REVIEW_PRESET, companyId };
if (base.rootIssueId && next.rootIssueId && base.rootIssueId !== next.rootIssueId) {
return deny(
"conflicting_low_trust_boundary",
source,
"Low-trust boundary root issue scopes do not overlap.",
sourcePresets,
);
}
return {
...base,
projectIds: intersectSets(base.projectIds, next.projectIds),
rootIssueId: base.rootIssueId ?? next.rootIssueId,
issueIds: intersectSets(base.issueIds, next.issueIds),
allowedAgentIds: intersectSets(base.allowedAgentIds, next.allowedAgentIds),
allowedSecretBindingIds: intersectSets(base.allowedSecretBindingIds, next.allowedSecretBindingIds),
allowedToolClasses: intersectSets(base.allowedToolClasses, next.allowedToolClasses),
outputPromotionTarget: next.outputPromotionTarget ?? base.outputPromotionTarget,
};
}
function hasBoundaryScope(boundary: LowTrustBoundary): boolean {
return Boolean(boundary.rootIssueId)
|| Boolean(boundary.projectIds?.length)
|| Boolean(boundary.issueIds?.length);
}
export function resolveCoreTrustPreset(input: ResolveCoreTrustPresetInput): TrustPresetResolution {
const sourcePresets: Partial<Record<TrustPresetPolicySource, TrustPreset>> = {};
const sources: ParsedPolicySource[] = [];
const agentPermissions = asRecord(input.agent?.permissions);
const agent = parseSource("agent", input.agent?.companyId, agentPermissions, agentPermissions?.authorizationPolicy, sourcePresets);
if ("kind" in agent) return agent;
sources.push(agent);
const projectPolicy = asRecord(input.project?.executionWorkspacePolicy);
const project = parseSource("project", input.project?.companyId, projectPolicy, projectPolicy?.authorizationPolicy, sourcePresets);
if ("kind" in project) return project;
sources.push(project);
const issuePolicy = asRecord(input.issue?.executionPolicy);
const issue = parseSource("issue", input.issue?.companyId, issuePolicy, issuePolicy?.authorizationPolicy, sourcePresets);
if ("kind" in issue) return issue;
sources.push(issue);
const runPolicy = asRecord(input.run?.executionPolicy);
const run = parseSource("run", input.run?.companyId, runPolicy, runPolicy?.authorizationPolicy, sourcePresets);
if ("kind" in run) return run;
sources.push(run);
for (const source of sources) {
if (source.companyId && source.companyId !== input.companyId) {
return deny(
"cross_company_boundary",
source.source,
"Policy source belongs to a different company.",
sourcePresets,
);
}
}
const effectivePreset = sources.some((source) => source.impliesLowTrust)
? LOW_TRUST_REVIEW_PRESET
: DEFAULT_TRUST_PRESET;
if (effectivePreset === DEFAULT_TRUST_PRESET) {
return {
kind: "standard",
preset: DEFAULT_TRUST_PRESET,
boundary: null,
sourcePresets,
};
}
let boundary: (LowTrustBoundary & { companyId: string }) | null = null;
for (const source of sources) {
if (!source.boundary) continue;
const merged = mergeBoundary(boundary, source.boundary, input.companyId, source.source, sourcePresets);
if (isTrustPresetResolution(merged)) return merged;
boundary = merged;
}
if (!boundary || !hasBoundaryScope(boundary)) {
return deny(
"missing_low_trust_boundary_scope",
null,
"Low-trust review requires a concrete project, root issue, or issue-id boundary.",
sourcePresets,
);
}
return {
kind: "low_trust_review",
preset: LOW_TRUST_REVIEW_PRESET,
boundary,
sourcePresets,
};
}
export function isIssueWithinLowTrustBoundary(
boundary: LowTrustBoundary & { companyId: string },
issue: { companyId: string; id?: string | null; projectId?: string | null },
): boolean {
if (issue.companyId !== boundary.companyId) return false;
if (issue.id && issue.id === boundary.rootIssueId) return true;
if (issue.id && boundary.issueIds?.includes(issue.id)) return true;
if (issue.projectId && boundary.projectIds?.includes(issue.projectId)) return true;
return false;
}