inspector/clients/web/src/utils/sandbox-csp.ts
Cliff Hall a82a70f4f5
refactor(web): sharpen the src/lib vs src/utils rule and move misfiled modules (#1784)
* refactor(web): sharpen src/lib vs src/utils rule and move misfiled modules (#1776)

Codify the empirical rule the web client already followed loosely — utils =
pure functions that compute; lib = things that instantiate, adapt, or touch
the environment — in AGENTS.md and clients/web/README.md, and move the modules
that sat on the wrong side of it.

Moves (source + co-located test):
- lib/sandbox-csp -> utils (pure CSP-string construction; type-only import)
- utils/browserTabVisibility -> lib (registers visibilitychange listeners)
- utils/oauthResume -> lib (reads/writes sessionStorage)
- utils/clearServerOAuthState -> lib (drives live InspectorClient / OAuthStorage)

Borderline audits (kept in place): pendingReauth (pure type declarations),
deepLink / correlateTransportErrors / mcpNetworkHeaders / oauthUx / oauthFlow /
oauthNetworkPhase (pure transforms / type-only core imports) stay in utils;
downloadFile stays whole in lib (dominant side is DOM side effects).

Also relocate the lone src/test/lib/downloadFile.test.ts straggler beside its
source at src/lib/downloadFile.test.ts, and document the test-placement rule
(side-by-side by default; src/test/ only for core/ tests, the integration
project, and shared scaffolding).

Pure move — no behavior change; no build/coverage config keys off the boundary.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5

* refactor(web): address review — fix stale paths, move customHeaders, one-way arrows (#1776)

Responds to the @claude review on PR #1784:

1. Stale path references (fixed): update the four in-repo pointers left at
   pre-move paths — specification/v2_auth_mid_session.md (oauthResume,
   browserTabVisibility → lib/), and the two load-bearing sandbox security
   comments (sandbox_proxy.html, sandbox-controller.ts → src/utils/sandbox-csp).

2. lib/types/customHeaders (moved): it's pure (types + pure transforms), the
   rule's clearest counterexample. Moved lib/types/customHeaders.{ts,test.ts} →
   utils/ (no external importers; only its own test) and removed the
   undocumented lib/types/ subdir. Documented that utils/ is the home for pure
   shared domain types + their constructors — no separate types/ bucket.

3. One-way arrows (item 3): declare OAuthResumeAuthKind in utils/pendingReauth
   and re-export it from lib/oauthResume, so the only cross-directory arrow now
   points lib → utils. Codified that invariant in AGENTS.md.

Wording nit: AGENTS.md + README now state diagnostic logging
(console.warn/error) doesn't count as a side effect for the utils rule.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5

* docs: address review round 2 — stale customHeaders refs, rule wording (#1776)

Doc-only follow-up to the round-2 @claude review on PR #1784 (no code changed):

1. Stale paths (same class as round-1 item 1): the customHeaders move left
   three references to the deleted lib/types/customHeaders.ts path in
   specification/ (v2_ux_interfaces_phase_0_4_inventory.md,
   v2_ux_interfaces_plan.md ×2) — updated to utils/customHeaders.ts.

2. customHeaders anchor: note in AGENTS.md that its shape is staged for
   ServerSettingsForm (per v2_ux_interfaces_plan.md), so its zero-importer
   state reads as intentional rather than rot; kept as the no-types/-bucket
   illustration and led the anchor list with load-bearing modules.

3. Type-only-import clause: split the conflated clause — a type-only import
   (pendingReauth) and re-exporting core's pure functions/constants
   (oauthUx/oauthFlow) are distinct reasons to stay in utils; what makes a
   module lib is wrapping core's stateful runtime. Fixed in both AGENTS.md
   and the web README.

Cosmetic: re-padded the v2_auth_mid_session.md table row to column width.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5

* docs: address review round 3 — scope the "no types/ bucket" claim (#1776)

Doc-only follow-up to the round-3 @claude review on PR #1784 (no code changed):

1. "no separate types/ bucket" overreached: clients/web/src/types/ does exist
   (navigation.ts + react-syntax-highlighter .d.ts ambient stubs wired via
   tsconfig.app.json paths). Scoped the claim in both AGENTS.md and the web
   README to "no types/ sub-bucket INSIDE lib/utils", and noted src/types/ is
   for ambient .d.ts stubs, not domain types.

2. README example list drifted from AGENTS.md's round-2 load-bearing anchors —
   realigned it (inspectorTabs/mcpNetworkHeaders, dropped the zero-importer
   customHeaders from the headline list).

3. Clarified v2_ux_interfaces_plan.md:185 so the v2 path reads as the copy's
   destination, not as the v1.5 source location.

Follow-up filed as #1785 (board Todo): src/types/navigation.ts InspectorTab is
dead and duplicates utils/inspectorTabs.ts — kept out of this pure-move PR.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5

* docs: address review round 4 — reconcile src/types/ claim with #1785 (#1776)

Doc-only follow-up to the round-4 @claude review on PR #1784 (no code changed):

The round-3 scope fix overshot: it asserted src/types/ holds "not domain types
at all", but the dead navigation.ts InspectorTab is a plain-.ts domain type
still living there, and specification/v2_ux_interfaces_plan.md:181-182 actively
prescribes placing InspectorTab there — so AGENTS.md and that spec doc gave
opposite guidance about the same directory.

- AGENTS.md + web README: softened to "not the place for new domain types"
  and added a parenthetical that the one lingering domain type (dead
  navigation.ts) is tracked for removal in #1785; until then a pure domain
  type belongs in utils/, not src/types/.
- v2_ux_interfaces_plan.md:181: marked the tab-enum placement guidance as
  superseded by #1785 (live type is InspectorTabId in utils/inspectorTabs.ts).

Self-heals on merge of #1785.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5

* docs: address review round 5 — fix garbled #1785 clause, restructure utils rule (#1776)

Doc-only follow-up to the round-5 @claude review on PR #1784 (no code changed):

1. Grammar (item 1): the round-4 #1785 clause read "superseding [#1785] will
   remove it" — a participle with no subject. Reworded to "the dead
   navigation.ts InspectorTab, which #1785 will remove".

2. Readability (item 2): the src/utils/ bullet had accreted into a single
   ~1,400-char blob across five review rounds. Promoted the carve-outs to
   scannable sub-bullets (Domain types / Diagnostic logging / Importing from
   core) and lifted the src/types/ note out to its own line, since it's a
   statement about a sibling directory, not about utils.

AGENTS.md only; the README's parallel sentence was already clean.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5

* docs+config: address review round 6 — correct the coverage-glob claim (#1776)

Responds to the round-6 @claude review on PR #1784:

1. "no build/coverage glob" was factually wrong (should-fix): the web coverage
   include in clients/web/vite.config.ts DOES enumerate src/lib/** and
   src/utils/**. The conclusion (a move between them is coverage-neutral)
   survives, but the reason was wrong. Reworded both AGENTS.md and the web
   README to say nothing *enforces* the boundary — no path alias, and the
   include lists both dirs so a move is coverage-neutral — plus the important
   whitelist caveat: a module outside components/lib/utils/server falls out of
   the ≥90 gate entirely.

2. Removed a pre-existing duplicate glob in vite.config.ts (src/lib/**/*.{ts,tsx}
   appeared twice, lines 161 and 163). Inert (vitest de-dupes) but it sat three
   lines from the config the doc now describes.

Nits: gave the README bullet the same round-5 sub-bullet restructure as
AGENTS.md, and made its #1785 reference a real markdown link (bare #N doesn't
autolink in rendered repo markdown).

npm run ci green (config change re-verified through coverage + smokes + Storybook).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01XNdjEPKLG637X8YmhDiEk5

---------

Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-25 22:24:30 -04:00

146 lines
6.4 KiB
TypeScript

import type { McpUiResourceCsp } from "@modelcontextprotocol/ext-apps/app-bridge";
/**
* Allowed shapes for a CSP source-expression supplied by an app's
* `_meta.ui.csp`. Each entry is server-supplied and untrusted: it MUST NOT
* inject extra directives (`;`) or break out of the meta attribute
* (`"`, `<`, `>`). Only common source forms are accepted —
* `scheme://host[:port][/path]`, scheme-only (`data:`, `blob:`), `*`, and
* wildcard hosts (`*.example.com`, `https://*.example.com`); anything else is
* dropped by {@link approveCspSources}.
*/
export const SAFE_CSP_SOURCE =
/^(?:\*|[a-zA-Z][a-zA-Z0-9+.-]*:(?:\/\/(?:\*\.)?[A-Za-z0-9._~%!$&'()*+,=@:-]+(?::\d+)?(?:\/[A-Za-z0-9._~%!$&'()*+,=@:/-]*)?)?|(?:\*\.)?[A-Za-z0-9](?:[A-Za-z0-9.-]*[A-Za-z0-9])?)$/;
/**
* Identity helper that pins `CSP_KEYS` to an exhaustive, valid key list at
* compile time. `satisfies readonly (keyof McpUiResourceCsp)[]` alone only
* proves every *listed* key is valid; it does NOT prove the list is *complete*.
* The intersected conditional adds the missing half: when `CSP_KEYS` covers
* every key, `keyof McpUiResourceCsp extends T[number]` holds and the parameter
* type is just `T`; if the upstream ext-apps type ever gains a new domain key,
* the conditional collapses the parameter type to `never` and the call fails to
* compile — forcing the key to be added here rather than being silently dropped
* (a vanished restriction) by {@link approveCspSources}.
*/
function exhaustiveCspKeys<const T extends readonly (keyof McpUiResourceCsp)[]>(
keys: T & (keyof McpUiResourceCsp extends T[number] ? unknown : never),
): T {
return keys;
}
const CSP_KEYS = exhaustiveCspKeys([
"connectDomains",
"resourceDomains",
"frameDomains",
"baseUriDomains",
]);
/**
* Filter an app-supplied {@link McpUiResourceCsp} down to the entries the host
* will actually enforce. Unsafe values are dropped (and warned), and the
* resulting object contains only keys with at least one accepted source. The
* return value is what the host echoes back to the view via
* `hostCapabilities.sandbox.csp` so the app sees what was granted, not what it
* asked for.
*
* NOTE: this screens each source for *injection safety* only — it does NOT
* bound the *breadth* of a grant. A bare `*` (and a scheme-wildcard host) is a
* syntactically safe source, so `resourceDomains: ["*"]` is "approved" and maps
* to `script-src 'unsafe-inline' *` (scripts from anywhere), just as
* `connectDomains: ["*"]` maps to `connect-src *`. That breadth is acceptable
* here because the enforced document runs in an opaque-origin sandbox with no
* ambient credentials and nothing to exfiltrate beyond what the app already
* received; "approved" therefore means "cannot break out of the meta
* attribute," not "restrictive."
*/
export function approveCspSources(
csp: McpUiResourceCsp | undefined,
): McpUiResourceCsp {
const approved: McpUiResourceCsp = {};
if (!csp) return approved;
for (const key of CSP_KEYS) {
const requested = csp[key];
if (!Array.isArray(requested)) continue;
const accepted: string[] = [];
for (const entry of requested) {
if (typeof entry === "string" && SAFE_CSP_SOURCE.test(entry)) {
accepted.push(entry);
} else {
console.warn("[mcp-app sandbox] dropping unsafe CSP source:", entry);
}
}
if (accepted.length > 0) approved[key] = accepted;
}
return approved;
}
function joinSources(list: string[] | undefined, fallback: string): string {
return list && list.length > 0 ? list.join(" ") : fallback;
}
/**
* Translate an approved {@link McpUiResourceCsp} into the Content-Security-Policy
* string enforced on the inner sandboxed document. `default-src 'none'` is the
* catch-all so any fetch type not explicitly mapped is denied. `script-src` /
* `style-src` carry `'unsafe-inline'` because the app's own inline code ships
* with the inline-delivered HTML and has no origin to allowlist; external loads
* stay restricted to `resourceDomains`.
*
* `resourceDomains` intentionally feeds `script-src` (and `style-src`) in
* addition to `img-src`/`font-src`/`media-src`: the `McpUiResourceCsp` contract
* defines it as a single "static resources" allowlist that "Maps to CSP
* `img-src`, `script-src`, `style-src`, `font-src`, `media-src` directives," so
* an app that lists a CDN there is granted script execution from that origin by
* design. There is no narrower per-directive key in the contract; if the spec
* ever splits scripts out, update the mapping here accordingly.
*/
export function buildSandboxCspPolicy(approved: McpUiResourceCsp): string {
const resourceSrc = joinSources(approved.resourceDomains, "'none'");
const inlineResource =
resourceSrc === "'none'"
? "'unsafe-inline'"
: `'unsafe-inline' ${resourceSrc}`;
return [
"default-src 'none'",
`connect-src ${joinSources(approved.connectDomains, "'none'")}`,
`script-src ${inlineResource}`,
`style-src ${inlineResource}`,
`img-src ${resourceSrc}`,
`font-src ${resourceSrc}`,
`media-src ${resourceSrc}`,
`frame-src ${joinSources(approved.frameDomains, "'none'")}`,
`base-uri ${joinSources(approved.baseUriDomains, "'self'")}`,
"form-action 'none'",
"object-src 'none'",
"worker-src 'none'",
].join("; ");
}
/** HTML-attribute-encode a string (defense-in-depth for the CSP meta value). */
export function escapeHtmlAttr(s: string): string {
return s
.replace(/&/g, "&amp;")
.replace(/"/g, "&quot;")
.replace(/'/g, "&#39;")
.replace(/</g, "&lt;")
.replace(/>/g, "&gt;");
}
/**
* Wrap an app's untrusted HTML in a host-authored document whose first
* `<head>` child is the CSP `<meta>`. The wrapper bytes are fixed — the
* untrusted content lands inside `<body>` and never precedes the policy, so a
* `<head>`/`<!-- -->` token in the app's HTML cannot push the meta inert or
* load resources before the policy applies. If the app's HTML is itself a full
* document, the second `<!doctype>`/`<html>`/`<head>` are parsed inside
* `<body>` (the HTML parser ignores duplicate document-structure tags) while
* its scripts and styles still run — governed by the already-applied policy.
*/
export function wrapSandboxedHtml(
untrustedHtml: string,
policy: string,
): string {
const meta = `<meta http-equiv="Content-Security-Policy" content="${escapeHtmlAttr(policy)}">`;
return `<!DOCTYPE html><html><head>${meta}<meta charset="utf-8"></head><body>${untrustedHtml}</body></html>`;
}