Files
PaperClipAI/packages/shared/src/mcp-skills.ts
T
DottaandClaude Opus 5.5 2d1c2db64a feat(mcp): host the MCP Skills extension for connected apps (#21219)
## Thinking Path

> - Paperclip is the open source app people use to manage AI agents for
work.
> - Agents use connected apps through MCP. Paperclip already hosts MCP
Apps views and speaks stateless MCP 2026-07-28.
> - The MCP Skills extension (SEP-2640, Final) lets an app publish
skills (instructions and files) next to its tools. Hugging Face already
serves 25.
> - Paperclip did not read them, so agents missed each app's own
workflow guidance.
> - Skills are a prompt-injection surface. The spec makes manifest
binding, verification, origin tagging and approval revocation host
requirements.
> - This pull request makes Paperclip a Skills host: discover, review,
verify, and deliver app skills to the agents that use the app.
> - The benefit is that connected apps can teach agents their workflows,
under board control.

## Linked Issues or Issue Description

**Problem**
Connected MCP servers can publish skills with the
`io.modelcontextprotocol/skills` extension. Paperclip ignored them. A
board could not see, review or approve them, and agents never received
them.

**Proposed solution**
Implement the host side of SEP-2640. List skills during catalog refresh,
record them per connection and company, bind board approval to the full
file manifest, and load approved skills into agent runs with verified,
lazily fetched content.

**Why it matters**
Apps ship workflow guidance with their tools. Without the extension,
each team must copy that guidance into company skills by hand, and it
goes stale.

Refs #21119, Refs #21140 (the MCP Apps and stateless MCP client work
this builds on).

## What Changed

- **Client capability.** Every stateless request declares
`io.modelcontextprotocol/skills`. Legacy handshake sessions do not. The
capabilities that `server/discover` advertises are remembered beside the
protocol era.
- **Discovery.** Catalog refresh calls `skills/list` (paged) only when
the server declared the extension. A listing inside its `ttlMs` is
reused. A skills failure never fails the tool refresh. Nothing runs
while a run starts.
- **Schema.** New `tool_mcp_skills` (company, connection, skill URI;
frontmatter, sorted manifest, manifest digest, review state) and
`tool_mcp_skill_files` (verified bytes keyed by connection, URI and
digest). Both cascade with the connection.
- **Governance.** New skills wait for review. Approval binds to the
order-free digest of every `{uri, digest, size}`; a request for any
other digest gets 409. A changed manifest or frontmatter revokes
approval and logs `tool_mcp_skill.approval_revoked`. Dynamic, oversized
and invalid skills are declined with a reason. Approve and decline use
the company skill policy (`skills.install`, `external_package`,
`mcp-connection:<id>`).
- **Run time.** Approved skills of connections the agent can use load in
parallel with other skill preparation, inside a 3 s budget. `SKILL.md`
is read with the agent's grant and checked for size, SHA-256 and
frontmatter. A failure calls `skills/get` and takes the skill out of
use. Each skill becomes a per-run bundle on the connector skill path,
named in its app's namespace, tagged with its origin. Paperclip writes
the frontmatter: `allowed-tools` only if the board approved it.
- **Supporting files.** `GET
/api/agents/me/mcp-skills/:skillId/files?path=` serves manifest-listed
files to an agent in an active run, through the skill's own connection.
The sandbox bridge allows the route.
- **Board API and UI.** `GET/POST /api/companies/:companyId/mcp-skills…`
(list, read instructions, approve, decline). The Skills page shows
"Skills from connected apps" with app, state, size, requested tools and
plain-language notes. It renders nothing when no app publishes skills.
- **Docs.** `doc/connections/MCP-SKILLS.md`, linked from
`doc/connections/README.md`.
- **Shared frontmatter parser.** It now decodes YAML single-quoted
scalars. Before, 5 of Hugging Face's 25 skills kept their quotes and
would fail the frontmatter check; now all 25 match. The generated runner
copy is regenerated.

File size notes: `server/src/services/tool-gateway.ts` grows by about
200 lines. The run-time reads must use its grant resolution and
`callAssignedConnectionProtocol`, like MCP Apps view reads. Splitting
the gateway is a separate refactor. `CompanySkills.tsx` and
`CompanySkills.production.tsx` gain 2–4 lines each; the section lives in
its own component.

## Verification

- `server/src/__tests__/mcp-skills.test.ts` (15 tests) uses a fake
stateless skills server and embedded Postgres. It covers: discovery only
after the server declares the extension; the capability in `_meta`;
`ttlMs` reuse; lazy fetch (no reads at listing or approval; `SKILL.md`
at load; supporting files on read); a stale-digest approval refused; a
reorder that keeps approval; a changed or added file that revokes it;
dynamic skills declined; a same-size tampered `SKILL.md` rejected,
refreshed with `skills/get`, and taken out of use; mismatched
frontmatter rejected; reads outside the manifest refused; the same URI
and name from two servers kept apart; company scoping; a slow app
skipped within the budget and cached later; HTTP authorization (board
only, agent needs an active run, run header must match) and activity
logging.
- Red/green: with the digest check disabled, the tamper test fails; with
the manifest comparison disabled, the revocation test fails. Both pass
with the code.
- `server/src/__tests__/mcp-protocol-versions.test.ts`,
`mcp-apps.test.ts`, `connection-instructions.test.ts`,
`tool-gateway-discovery-http.test.ts`: 47 pass (two capability
assertions updated for the new extension).
- `packages/shared/src/frontmatter.test.ts` adds a single-quoted scalar
case; the runner contract test passes (23 tests).
- `ui/src/components/skills/ConnectedAppSkills.test.tsx` (3) and
`ui/src/pages/CompanySkills.test.tsx` (44) pass.
`packages/adapter-utils/src/sandbox-callback-bridge.test.ts` (89)
passes.
- `pnpm --filter @paperclipai/server typecheck` passes. UI `tsc` passes
except the known missing generated Storybook catalog files.
- `pnpm verify` returned exit 2 (doc change is a global CI input);
required CI supplies broad coverage.
- Live check against `https://huggingface.co/mcp` without credentials,
through Paperclip's own services and the real egress guard: 25 skills
discovered as pending; three approved; cold load 135 ms, cached load 13
ms; a supporting file read through the manifest; the bundle tagged with
its origin.
- Storybook: `Agents and reusable capabilities/Skills and Skill
Studio/Connected app skills` (needs review, all states, tool request,
name collision, reading instructions).

## Risks

- **Migration:** two new tables only. No changes to existing tables.
Safe with the previous release running.
- **Run start:** companies without an approved app skill pay one indexed
read. With approved skills the wait is capped at 3 s; late or failing
skills are left out with a `skills.mcp_unavailable` run warning and
never fail the run.
- **Agent harness enforcement:** Paperclip cannot gate each
code-execution tool call an agent makes while following an app's skill.
Approval is explicit and per skill, the approval note says agents may
run commands the skill describes, and `allowed-tools` is never granted
without a separate tick.
- **Shared-home adapters** receive app skills like company skills. A
revoked skill leaves on the next skill sync.
- Not in this PR: per-connection opt-in for dynamic skills; serving
Paperclip's own skills from its public MCP endpoint;
`resources/directory/read`; app skills on an agent's skills tab.

## Model Used

- Claude Opus 5.5 (`claude-opus-5-5`), Anthropic, via Claude Code with
tool use (shell, file edits, local test runs, live HTTP checks).
Extended reasoning 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] I have not referenced internal/instance-local Paperclip issues or
links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip`
URLs)
- [x] My branch name describes the change (e.g. `docs/...`, `fix/...`)
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

🤖 Generated with [Claude Code](https://claude.com/claude-code)

---------

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
2026-11-12 19:18:06 -06:00

171 lines
7.6 KiB
TypeScript

/**
* MCP Skills extension (SEP-2640): skills that a connected MCP server
* publishes alongside its tools. Paperclip is the host. It lists each server's
* skills, asks the board to approve each one against its exact file manifest,
* and delivers approved skills to agents that can use the connection.
*
* https://github.com/modelcontextprotocol/ext-skills/blob/main/specification/beta/skills.mdx
*/
import { z } from "zod";
export const MCP_SKILLS_EXTENSION_ID = "io.modelcontextprotocol/skills";
/** Client capability Paperclip declares in each request's `_meta`. */
export const MCP_SKILLS_CLIENT_CAPABILITY = { [MCP_SKILLS_EXTENSION_ID]: {} } as const;
/** Per-skill limits every conforming host accepts (spec: Limits). */
export const MCP_SKILL_MAX_RESOURCES = 512;
export const MCP_SKILL_MAX_TOTAL_BYTES = 16 * 1024 * 1024;
export interface McpSkillResource {
uri: string;
/** `sha256:` followed by 64 lowercase hex characters. */
digest: string;
size: number;
}
export interface McpSkillEntry {
uri: string;
frontmatter: { name: string; description: string; [key: string]: unknown };
resources: McpSkillResource[] | "dynamic";
}
export const MCP_SKILL_STATUSES = ["pending_review", "approved", "declined", "unavailable"] as const;
export type McpSkillStatus = (typeof MCP_SKILL_STATUSES)[number];
export const MCP_SKILL_STATUS_REASONS = [
"new",
"changed",
"dynamic",
"too_large",
"invalid",
"verification_failed",
"removed",
] as const;
export type McpSkillStatusReason = (typeof MCP_SKILL_STATUS_REASONS)[number];
/** A connected app's skill as the board sees it. */
export interface CompanyMcpSkill {
id: string;
connectionId: string;
appName: string;
name: string;
description: string;
skillUri: string;
status: McpSkillStatus;
statusReason: McpSkillStatusReason | null;
/** The digest the board approves; the server rejects an approval for any other. */
manifestDigest: string | null;
fileCount: number | null;
totalBytes: number | null;
/** The tools the skill asks to use (`allowed-tools`). Never granted without approval. */
requestedTools: string[];
allowedToolsApproved: boolean;
/** Other skills, from this or another source, with the same name. */
nameCollision: boolean;
reviewedAt: string | null;
lastSeenAt: string;
}
export const approveCompanyMcpSkillSchema = z.object({
/** The manifest digest the board reviewed. */
manifestDigest: z.string().regex(/^sha256:[0-9a-f]{64}$/),
/** Also honor the skill's `allowed-tools`. */
allowTools: z.boolean().optional(),
}).strict();
export type ApproveCompanyMcpSkillRequest = z.infer<typeof approveCompanyMcpSkillSchema>;
export const mcpSkillFileQuerySchema = z.object({ path: z.string().min(1).max(1024) });
const DIGEST_RE = /^sha256:[0-9a-f]{64}$/;
// Agent Skills spec: 1-64 chars, lowercase letters, digits and single hyphens.
const SKILL_NAME_RE = /^[a-z0-9](?:[a-z0-9]|-(?!-)){0,63}$/;
function record(value: unknown): Record<string, unknown> | null {
return value && typeof value === "object" && !Array.isArray(value) ? value as Record<string, unknown> : null;
}
/** The skill root (the SKILL.md URI without `/SKILL.md`), or null when the URI is not a skill URI. */
export function mcpSkillRoot(skillUri: string): string | null {
return skillUri.endsWith("/SKILL.md") ? skillUri.slice(0, -"/SKILL.md".length) : null;
}
/** The skill's name as recoverable from its URI: the last segment of the skill path. */
export function mcpSkillNameFromUri(skillUri: string): string | null {
const root = mcpSkillRoot(skillUri);
if (!root) return null;
const segment = root.slice(root.lastIndexOf("/") + 1);
return segment || null;
}
/** A file path relative to the skill root, or null when the URI is outside the skill. */
export function mcpSkillRelativePath(skillUri: string, fileUri: string): string | null {
const root = mcpSkillRoot(skillUri);
if (!root || !fileUri.startsWith(`${root}/`)) return null;
const relative = fileUri.slice(root.length + 1);
const segments = relative.split("/");
if (!relative || segments.some((segment) => !segment || segment === "." || segment === "..")) return null;
return relative;
}
export type McpSkillEntryCheck =
| { ok: true; entry: McpSkillEntry; totalBytes: number | null }
| { ok: false; uri: string | null; reason: "invalid" | "too_large"; message: string };
/**
* Validate one `skills/list` or `skills/get` entry against the spec's
* structural rules. An invalid entry is never loaded; one over the limits is
* declined with a reason the board can see.
*/
export function checkMcpSkillEntry(value: unknown): McpSkillEntryCheck {
const entry = record(value);
const uri = typeof entry?.uri === "string" ? entry.uri : null;
const invalid = (message: string): McpSkillEntryCheck => ({ ok: false, uri, reason: "invalid", message });
if (!entry || !uri || uri.length > 2048) return invalid("The skill has no valid URI.");
const frontmatter = record(entry.frontmatter);
if (!frontmatter || typeof frontmatter.name !== "string" || typeof frontmatter.description !== "string" || !frontmatter.description.trim()) {
return invalid("The skill's frontmatter must include a name and a description.");
}
if (!SKILL_NAME_RE.test(frontmatter.name)) return invalid("The skill's name is not a valid skill name.");
if (mcpSkillNameFromUri(uri) !== frontmatter.name) return invalid("The skill's URI does not end in its name.");
const resources = entry.resources;
if (resources === "dynamic") {
return { ok: true, entry: { uri, frontmatter: frontmatter as McpSkillEntry["frontmatter"], resources }, totalBytes: null };
}
if (!Array.isArray(resources)) return invalid("The skill has no file manifest.");
if (resources.length > MCP_SKILL_MAX_RESOURCES) {
return { ok: false, uri, reason: "too_large", message: `The skill has more than ${MCP_SKILL_MAX_RESOURCES} files.` };
}
const seen = new Set<string>();
const files: McpSkillResource[] = [];
let totalBytes = 0;
for (const item of resources) {
const resource = record(item);
const fileUri = typeof resource?.uri === "string" ? resource.uri : null;
const digest = typeof resource?.digest === "string" ? resource.digest : null;
const size = resource?.size;
if (!fileUri || !digest || !DIGEST_RE.test(digest) || typeof size !== "number" || !Number.isSafeInteger(size) || size < 0) {
return invalid("A file in the skill's manifest has no valid URI, digest or size.");
}
if (fileUri !== uri && !mcpSkillRelativePath(uri, fileUri)) return invalid("The skill's manifest lists a file outside the skill.");
if (seen.has(fileUri)) return invalid("The skill's manifest lists a file twice.");
seen.add(fileUri);
totalBytes += size;
files.push({ uri: fileUri, digest, size });
}
if (!seen.has(uri)) return invalid("The skill's manifest does not include its SKILL.md.");
if (totalBytes > MCP_SKILL_MAX_TOTAL_BYTES) {
return { ok: false, uri, reason: "too_large", message: "The skill's files are larger than 16 MiB in total." };
}
files.sort((a, b) => (a.uri < b.uri ? -1 : a.uri > b.uri ? 1 : 0));
return { ok: true, entry: { uri, frontmatter: frontmatter as McpSkillEntry["frontmatter"], resources: files }, totalBytes };
}
/** The `allowed-tools` a skill requests, in either the list or the space-separated form. */
export function mcpSkillRequestedTools(frontmatter: Record<string, unknown>): string[] {
const value = frontmatter["allowed-tools"];
const tools = Array.isArray(value) ? value : typeof value === "string" ? value.split(/\s+/) : [];
return tools.filter((tool): tool is string => typeof tool === "string" && tool.trim().length > 0).map((tool) => tool.trim());
}