Files
PaperClipAI/ui/src/pages/tools/profiles/profile-model.ts
T
DottaandPaperclip a116713b9a feat(mcp) [split 6/8]: add Tools and Profiles UI foundation (#9561)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work
> - Governed MCP access spans contracts, runtime enforcement, adapters,
UI surfaces, and operator verification
> - The parity reference PR #9534 is too large for effective automated
or human review
> - The feature therefore needs a linear stack whose individual diffs
stay below the 100-file review limit
> - This pull request is split 6/8 and focuses on UI API, shared
components, and Tools/Profile surfaces
> - The benefit is a standalone, testable review boundary while
preserving byte-for-byte parity at the top of the stack

## Linked Issues or Issue Description

- Related parity reference: #9534
- Problem: Operators need typed clients and administration surfaces that
compile independently before navigation exposes them.
- Proposed solution: Adds UI APIs, hooks, libraries, shared components,
Tools/Profiles pages, and the plugin settings consumer required by the
new company-scoped API.
- Alternatives considered: keeping #9534 as one 403-file review, or
rewriting the feature to manufacture seams; both were rejected in favor
of path extraction plus compile-driven boundary moves.
- Roadmap alignment: this advances the existing governed MCP/tool-access
work already represented by #9534; it does not introduce a separate
roadmap initiative.
- Stack position: base branch is
`pap10341-split/05-runtime-integration`.
- Merge policy: merge bottom-up, in order, only after the complete
eight-PR stack has been reviewed and the top-of-stack parity gate
remains empty.
- Requested review: UXDesigner sanity pass on Tools/Profile surfaces;
Greptile on every PR.

## What Changed

- Adds UI APIs, hooks, libraries, shared components, Tools/Profiles
pages, and the plugin settings consumer required by the new
company-scoped API.
- Keeps this PR below 100 changed files and independently typecheckable.
- Preserves the final tree from #9534 when combined with the other seven
stack levels.

## Verification

- `pnpm typecheck`
- `pnpm check:token-gates` — all gates clean
- Focused UI Vitest run with `NODE_ENV=test` — 28 files, 183 tests
passed

## Risks

- Large dead-code UI additions can drift from activation routes; PR 7
supplies the registration layer and top-level parity catches omissions.
- Stack risk: merging out of order can expose incomplete layers;
mitigate by following the documented bottom-up merge policy.
- Parity risk: later edits to an intermediate branch can drift from
#9534; mitigate by re-running the empty top-of-stack diff before merge.

> 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, exact model ID `gpt-5.4`; runtime-managed context
window; medium reasoning with repository, shell, Git, GitHub CLI, and
code-execution tools enabled.

## 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] Internal references are omitted except the execution-plan link
explicitly required for this coordinated split stack
- [x] My branch name describes the change 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
- [ ] All Paperclip CI gates are green
- [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups
- [x] I will address all Greptile and reviewer comments before
requesting merge


## Stack Coordination

- Internal execution plan:
[PAP-13874](/PAP/issues/PAP-13874#document-plan)
- Parity reference: #9534
- Stack: #9556 → #9557 → #9558 → #9559 → #9560 → #9561 → #9562 → #9563
- Merge bottom-up only after full-stack review and an empty parity diff
at #9563.


## UI Evidence

QA captured these from the live Garden MCP split stack at 1440px and
verified clean rendering:

![Profiles
index](https://raw.githubusercontent.com/paperclipai/paperclip/ac72de3ec85feb14c5ec1ed0a3d39b29baa0abff/doc/screenshots/garden-mcp-split-stack/profiles-index.png)

![Profile
wizard](https://raw.githubusercontent.com/paperclipai/paperclip/ac72de3ec85feb14c5ec1ed0a3d39b29baa0abff/doc/screenshots/garden-mcp-split-stack/profile-wizard.png)


![Policies](https://raw.githubusercontent.com/paperclipai/paperclip/ac72de3ec85feb14c5ec1ed0a3d39b29baa0abff/doc/screenshots/garden-mcp-split-stack/policies.png)

---------

Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-07-14 15:31:12 -05:00

400 lines
16 KiB
TypeScript

import type {
ToolCatalogEntry,
ToolProfileDefaultAction,
ToolProfileEntry,
ToolRiskLevel,
} from "@paperclipai/shared";
import type { ToolProfileEntryInput } from "@/api/tools";
/**
* Pure domain model for the prosumer access-profile wizard (PAP-10997).
*
* The wizard speaks a friendly per-app language ("All Gmail tools", "All Gmail
* except 2", "3 Gmail tools") while the server stores the existing entry model
* (`application` / `catalog_entry` selectors with include/exclude effects). This
* module is the single source of truth for that translation and for the
* "except N" math, so it can be unit-tested without React.
*
* Vocabulary gate: nothing here renders. Labels produced here are prosumer copy
* ("tools", "All … except N"); selector/entry/effect/default-action vocabulary
* never leaves this module.
*/
// --- Capability ------------------------------------------------------------
export type ToolCapability = "read" | "write" | "destructive";
export function toolCapability(tool: ToolCatalogEntry): ToolCapability {
if (tool.isDestructive) return "destructive";
if (tool.isWrite) return "write";
return "read";
}
export const CAPABILITY_LABEL: Record<ToolCapability, string> = {
read: "Read-only",
write: "Makes changes",
destructive: "Destructive",
};
// --- App grouping ----------------------------------------------------------
/** A connected app and the concrete tools its catalog currently exposes. */
export interface AppGroup {
/** Stable grouping key — the application id, or the connection id when an
* entry has no application (admin "run your own" connections). */
appKey: string;
applicationId: string | null;
connectionId: string;
name: string;
tools: ToolCatalogEntry[];
}
/**
* Group the company-wide catalog into apps, preferring the application name and
* falling back to the connection name for application-less connections.
*/
export function groupCatalogByApp(
catalog: ToolCatalogEntry[],
applicationsById: Map<string, string>,
connectionsById: Map<string, string>,
): AppGroup[] {
const groups = new Map<string, AppGroup>();
for (const tool of catalog) {
const appKey = tool.applicationId ?? tool.connectionId;
let group = groups.get(appKey);
if (!group) {
const name =
(tool.applicationId ? applicationsById.get(tool.applicationId) : null) ??
connectionsById.get(tool.connectionId) ??
"Tools";
group = {
appKey,
applicationId: tool.applicationId,
connectionId: tool.connectionId,
name,
tools: [],
};
groups.set(appKey, group);
}
group.tools.push(tool);
}
for (const group of groups.values()) {
group.tools.sort((a, b) => a.toolName.localeCompare(b.toolName));
}
return [...groups.values()].sort((a, b) => a.name.localeCompare(b.name));
}
// --- Per-app selection state ----------------------------------------------
export type AppSelection =
| { kind: "all" }
| { kind: "all_except"; excluded: string[] }
| { kind: "some"; included: string[] }
| { kind: "none" };
export type WizardSelections = Record<string, AppSelection>;
/** Set of catalog-entry ids currently allowed for one app, given the catalog. */
export function selectedToolIds(group: AppGroup, selection: AppSelection | undefined): Set<string> {
if (!selection || selection.kind === "none") return new Set();
if (selection.kind === "all") return new Set(group.tools.map((t) => t.id));
if (selection.kind === "all_except") {
const excluded = new Set(selection.excluded);
return new Set(group.tools.filter((t) => !excluded.has(t.id)).map((t) => t.id));
}
const included = new Set(selection.included);
return new Set(group.tools.filter((t) => included.has(t.id)).map((t) => t.id));
}
export function isToolSelected(
group: AppGroup,
selection: AppSelection | undefined,
toolId: string,
): boolean {
return selectedToolIds(group, selection).has(toolId);
}
export type AppCheckState = "checked" | "indeterminate" | "unchecked";
export function appCheckState(group: AppGroup, selection: AppSelection | undefined): AppCheckState {
const n = selectedToolIds(group, selection).size;
if (n === 0) return "unchecked";
if (n === group.tools.length) return "checked";
return "indeterminate";
}
/**
* "All Gmail tools (12)" / "All Gmail except 2" / "3 of 12 tools" — the prosumer
* summary the tree row and the index render. `appName` is woven in by the caller
* when it wants the app-scoped variant.
*/
export function appSelectionLabel(group: AppGroup, selection: AppSelection | undefined): string {
const total = group.tools.length;
const state = appCheckState(group, selection);
if (state === "unchecked") return "None selected";
if (selection?.kind === "all" || (selection?.kind === "all_except" && selection.excluded.length === 0)) {
return `All ${group.name} tools (${total})`;
}
if (selection?.kind === "all_except") {
const n = selection.excluded.length;
return `All ${group.name} except ${n}`;
}
const n = selectedToolIds(group, selection).size;
return `${n} of ${total} ${group.name} tools`;
}
// --- Checkbox reducers -----------------------------------------------------
/** Toggle the app-level checkbox: off → all tools; on → none. */
export function toggleApp(group: AppGroup, selection: AppSelection | undefined): AppSelection {
const state = appCheckState(group, selection);
return state === "unchecked" ? { kind: "all" } : { kind: "none" };
}
/** Toggle one tool, preserving "include future tools" intent when the app box is on. */
export function toggleTool(
group: AppGroup,
selection: AppSelection | undefined,
toolId: string,
): AppSelection {
const current = selection ?? { kind: "none" as const };
const selected = selectedToolIds(group, current);
const willSelect = !selected.has(toolId);
// App box is on (all / all_except): keep app-include semantics, edit excludes.
if (current.kind === "all" || current.kind === "all_except") {
const excluded = new Set(current.kind === "all_except" ? current.excluded : []);
if (willSelect) excluded.delete(toolId);
else excluded.add(toolId);
if (excluded.size === 0) return { kind: "all" };
if (excluded.size >= group.tools.length) return { kind: "none" };
return { kind: "all_except", excluded: group.tools.filter((t) => excluded.has(t.id)).map((t) => t.id) };
}
// App box is off: explicit per-tool includes (future tools stay blocked).
const included = new Set(current.kind === "some" ? current.included : []);
if (willSelect) included.add(toolId);
else included.delete(toolId);
if (included.size === 0) return { kind: "none" };
return { kind: "some", included: group.tools.filter((t) => included.has(t.id)).map((t) => t.id) };
}
// --- Advanced rules (AP18) -------------------------------------------------
export type AdvancedRuleKind = "tool_name" | "risk_level" | "catalog_entry";
export interface AdvancedRule {
id: string;
kind: AdvancedRuleKind;
/** wildcard pattern for tool_name, a risk level, or a catalog id. */
value: string;
riskLevel?: ToolRiskLevel;
effect: "include" | "exclude";
}
// --- Entry <-> selection translation --------------------------------------
/**
* Flatten the wizard's per-app selections (+ advanced rules) into the server
* entry model. App-on selections become an `application` include; per-tool
* opt-outs become `catalog_entry` excludes; explicit picks become
* `catalog_entry` includes.
*/
export function buildEntries(
groups: AppGroup[],
selections: WizardSelections,
advancedRules: AdvancedRule[] = [],
defaultAction: ToolProfileDefaultAction = "deny",
): ToolProfileEntryInput[] {
const entries: ToolProfileEntryInput[] = [];
for (const group of groups) {
const selection = selections[group.appKey];
if (!selection) continue;
if (defaultAction === "allow") {
const selectedIds = selectedToolIds(group, selection);
for (const tool of group.tools) {
if (!selectedIds.has(tool.id)) {
entries.push({ selectorType: "catalog_entry", effect: "exclude", catalogEntryId: tool.id });
}
}
continue;
}
if (selection.kind === "none") continue;
if (selection.kind === "all" || selection.kind === "all_except") {
if (group.applicationId) {
entries.push({ selectorType: "application", effect: "include", applicationId: group.applicationId });
} else {
entries.push({ selectorType: "connection", effect: "include", connectionId: group.connectionId });
}
if (selection.kind === "all_except") {
for (const toolId of selection.excluded) {
entries.push({ selectorType: "catalog_entry", effect: "exclude", catalogEntryId: toolId });
}
}
} else {
for (const toolId of selection.included) {
entries.push({ selectorType: "catalog_entry", effect: "include", catalogEntryId: toolId });
}
}
}
for (const rule of advancedRules) {
if (rule.kind === "tool_name") {
entries.push({ selectorType: "tool_name", effect: rule.effect, toolName: rule.value });
} else if (rule.kind === "risk_level") {
entries.push({ selectorType: "risk_level", effect: rule.effect, riskLevel: rule.riskLevel ?? "destructive" });
} else {
entries.push({ selectorType: "catalog_entry", effect: rule.effect, catalogEntryId: rule.value });
}
}
return entries;
}
/**
* Recover wizard state from stored entries (draft resume / copy a profile).
* App-include + catalog excludes round-trip back to all / all_except; bare
* catalog includes become explicit picks; everything else (wildcards, risk
* levels, cross-app catalog ids) surfaces as an advanced rule.
*/
export function parseEntries(
groups: AppGroup[],
entries: ToolProfileEntry[],
): { selections: WizardSelections; advancedRules: AdvancedRule[] } {
const selections: WizardSelections = {};
const advancedRules: AdvancedRule[] = [];
const toolIdToApp = new Map<string, string>();
for (const group of groups) {
for (const tool of group.tools) toolIdToApp.set(tool.id, group.appKey);
}
const appByApplicationId = new Map<string, AppGroup>();
const appByConnectionId = new Map<string, AppGroup>();
for (const group of groups) {
if (group.applicationId) appByApplicationId.set(group.applicationId, group);
appByConnectionId.set(group.connectionId, group);
}
const appOn = new Set<string>();
const excludesByApp = new Map<string, Set<string>>();
const includesByApp = new Map<string, Set<string>>();
for (const entry of entries) {
if (entry.effect === "include" && entry.selectorType === "application" && entry.applicationId) {
const group = appByApplicationId.get(entry.applicationId);
if (group) appOn.add(group.appKey);
continue;
}
if (entry.effect === "include" && entry.selectorType === "connection" && entry.connectionId) {
const group = appByConnectionId.get(entry.connectionId);
if (group && !group.applicationId) appOn.add(group.appKey);
else advancedRules.push({ id: entry.id, kind: "catalog_entry", value: entry.connectionId, effect: "include" });
continue;
}
if (entry.selectorType === "catalog_entry" && entry.catalogEntryId) {
const appKey = toolIdToApp.get(entry.catalogEntryId);
if (!appKey) {
advancedRules.push({ id: entry.id, kind: "catalog_entry", value: entry.catalogEntryId, effect: entry.effect });
continue;
}
const bucket = entry.effect === "exclude" ? excludesByApp : includesByApp;
const set = bucket.get(appKey) ?? new Set<string>();
set.add(entry.catalogEntryId);
bucket.set(appKey, set);
continue;
}
if (entry.selectorType === "tool_name" && entry.toolName) {
advancedRules.push({ id: entry.id, kind: "tool_name", value: entry.toolName, effect: entry.effect });
continue;
}
if (entry.selectorType === "risk_level" && entry.riskLevel) {
advancedRules.push({ id: entry.id, kind: "risk_level", value: entry.riskLevel, riskLevel: entry.riskLevel, effect: entry.effect });
continue;
}
}
for (const group of groups) {
const on = appOn.has(group.appKey);
const excluded = [...(excludesByApp.get(group.appKey) ?? [])].filter((id) =>
group.tools.some((t) => t.id === id),
);
const included = [...(includesByApp.get(group.appKey) ?? [])].filter((id) =>
group.tools.some((t) => t.id === id),
);
if (on) {
selections[group.appKey] = excluded.length > 0 ? { kind: "all_except", excluded } : { kind: "all" };
} else if (included.length > 0) {
selections[group.appKey] = { kind: "some", included };
} else {
selections[group.appKey] = { kind: "none" };
}
}
return { selections, advancedRules };
}
// --- Live count ------------------------------------------------------------
/** "Allows 14 of 63 tools" — resolves the assembled selection against the catalog. */
export function countAllowedTools(
groups: AppGroup[],
selections: WizardSelections,
defaultAction: ToolProfileDefaultAction,
totalToolCount: number,
): { allowed: number; total: number } {
if (defaultAction === "allow") {
// "Allowed automatically" — everything except the per-app opt-outs.
let excluded = 0;
for (const group of groups) {
const selection = selections[group.appKey];
if (selection?.kind === "all_except") excluded += selection.excluded.length;
else if (selection?.kind === "some") excluded += group.tools.length - selection.included.length;
else if (selection?.kind === "none") excluded += group.tools.length;
}
return { allowed: Math.max(0, totalToolCount - excluded), total: totalToolCount };
}
let allowed = 0;
for (const group of groups) {
allowed += selectedToolIds(group, selections[group.appKey]).size;
}
return { allowed, total: totalToolCount };
}
// --- Step-1 templates ------------------------------------------------------
export type TemplateKey = "read_only" | "everyday" | "full_access" | "scratch" | "copy";
export interface TemplateDef {
key: TemplateKey;
title: string;
description: string;
}
export const TEMPLATES: TemplateDef[] = [
{ key: "read_only", title: "Read-only", description: "See and fetch, but never change anything." },
{ key: "everyday", title: "Everyday work", description: "Read and make routine changes — no destructive tools." },
{ key: "full_access", title: "Full access", description: "Everything every connected app offers." },
{ key: "scratch", title: "Start from scratch", description: "An empty profile you build up tool by tool." },
{ key: "copy", title: "Copy an existing profile", description: "Start from a profile you already have." },
];
function capabilityPredicate(key: TemplateKey): (tool: ToolCatalogEntry) => boolean {
if (key === "read_only") return (t) => toolCapability(t) === "read";
if (key === "everyday") return (t) => toolCapability(t) !== "destructive";
return () => true; // full_access
}
/** Build the initial per-app selection a template implies for the live catalog. */
export function templateSelections(key: TemplateKey, groups: AppGroup[]): WizardSelections {
const selections: WizardSelections = {};
if (key === "scratch" || key === "copy") {
for (const group of groups) selections[group.appKey] = { kind: "none" };
return selections;
}
const matches = capabilityPredicate(key);
for (const group of groups) {
const matchingIds = group.tools.filter(matches).map((t) => t.id);
if (matchingIds.length === 0) selections[group.appKey] = { kind: "none" };
else if (matchingIds.length === group.tools.length) selections[group.appKey] = { kind: "all" };
else selections[group.appKey] = { kind: "some", included: matchingIds };
}
return selections;
}