foldrun 10.3.1 → 10.4.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/bin/foldrun.mjs CHANGED
@@ -170,6 +170,8 @@ const HELP = `foldrun — agents are just folders
170
170
  foldrun inbox what waits on you: approvals, questions, issues, proposed fixes, your failed runs (--team, --all, --kind, --unread, --dismissed)
171
171
  foldrun inbox read|unread|dismiss|restore <id>… mark items in your own view (dismiss sets aside; it still waits where it lives)
172
172
  foldrun mcp this account as MCP tools for an AI app, over stdio, with this machine's sign-in: \`claude mcp add foldrun -- foldrun mcp\` (--read-only: only the tools that read)
173
+ foldrun mcp connections the AI apps connected over OAuth (claude.ai and the like): which app, whose, what access, last used — everyone's for an admin, else yours (--all adds the disconnected, --json)
174
+ foldrun mcp disconnect <id> end one: every token it holds stops working (asks; --yes)
173
175
  foldrun issues what went wrong, grouped, open until resolved (--to <workspace>, --all, --status, --json)
174
176
  foldrun issues show|ack|snooze|resolve|reopen <id> one issue — snooze --for 1d or --times 5, --note "…"
175
177
  foldrun issues doctor <id> the run doctor on the issue's newest run, in its workspace: why, and the fix as diffs, saved on yes (--yes, --dry-run, --json)
@@ -351,7 +353,7 @@ Options
351
353
  --value "<text>" secrets set: skip the prompt (careful with shell history)
352
354
  --site <domain> secrets set, secrets sites: a site the secret is bound to — the egress proxy fills it for those alone (repeatable, or a,b)
353
355
  --clear secrets sites: unbind it; flow compute: no cap of its own
354
- --all secrets workspaces: every workspace may use the account secret again
356
+ --all secrets workspaces: every workspace may use the account secret again; mcp connections: the disconnected ones too
355
357
  --none secrets workspaces, secrets set: no workspace may use the account secret
356
358
  --workspaces <a,b> secrets set: store an account secret limited to those workspaces, in the same write (with --account)
357
359
  --add <site> agent network: add a site to its network: list (repeatable)
package/docs/api.md CHANGED
@@ -17,7 +17,7 @@ proof instead of a key: a token or signature (`/s/<token>`, `/api/hooks`,
17
17
  `/api/inbox`, `/api/events`, `/api/approve`, `/api/unsubscribe`, `/api/billing/stripe`,
18
18
  `/api/billing/confirm`), the sign-in and recovery steps (`/api/auth/signup`,
19
19
  `login`, `forgot`, `reset`, `verify`, `mfa/verify`, `session`, `/api/cli/login`,
20
- `/api/cli/login/<id>`, `/api/cli/authorize` GET, `/api/oauth/callback`), `/api/healthz`, `/api/status`, `/api/status/history`, `/api/version` and `/api/openapi.json`.
20
+ `/api/cli/login/<id>`, `/api/cli/authorize` GET, `/api/oauth/callback`), `/api/healthz`, `/api/status`, `/api/status/history`, `/api/version` and `/api/openapi.json` — and an AI app's OAuth sign-in to `/api/mcp`: `/api/mcp/oauth/register` (open, as the protocol needs) and `/api/mcp/oauth/token` (its code and PKCE verifier, or its refresh token, are the proof).
21
21
 
22
22
  ```
23
23
  Authorization: Bearer <api-key> # machines. Settings → API keys
@@ -528,12 +528,12 @@ published, is `404` like one that does not exist.
528
528
 
529
529
  ## MCP — AI apps connecting with a key
530
530
 
531
- The platform as an [MCP server](mcp) for Claude Code, Claude Desktop,
532
- Cursor, Codex and any other MCP client. An **API key** opens it
533
- (`Authorization: Bearer mda_…`) — never a session cookie, never an A2A key —
534
- and every tool is a call to one of the routes above, made with that key, so
535
- it answers what the route would: the key's role, its workspaces and every
536
- refusal hold. MCP's Streamable HTTP transport, stateless: one JSON-RPC 2.0
531
+ The platform as an [MCP server](mcp) for claude.ai, Claude Code, Claude
532
+ Desktop, Cursor, Codex and any other MCP client. An **API key**
533
+ (`Authorization: Bearer mda_…`) or an AI app's **OAuth access token** (below)
534
+ opens it — never a session cookie, never an A2A key — and every tool is a call
535
+ to one of the routes above, made with that credential, so it answers what the
536
+ route would: its role, its workspaces and every refusal hold. MCP's Streamable HTTP transport, stateless: one JSON-RPC 2.0
537
537
  message per request, answered as JSON — no SSE stream, no session id, no
538
538
  batches. Errors before a message is read (no key, a refused key, a foreign
539
539
  `Origin`, an `MCP-Protocol-Version` it does not speak) are JSON-RPC error
@@ -543,3 +543,22 @@ with `HTTP <status>: <the route's error>`.
543
543
  | Route | Methods | |
544
544
  |---|---|---|
545
545
  | `/api/mcp` | POST | one JSON-RPC message: `initialize` (revisions `2025-11-25`, `2025-06-18`, `2025-03-26`; an unknown one is answered with the newest), `ping`, `tools/list`, `tools/call`; a notification is `202` with no body. `?readonly=1` offers only the tools that read and refuses the rest (`-32602`). The tools: `whoami`, `list_workspaces`, `inbox`, `list_approvals`, `list_flows`, `list_runs`, `get_run`, `read_file`, `run_flow`, `decide_approval`, `answer_question`, `message_run`, `stop_run` — see [MCP](mcp#the-tools). `foldrun mcp` serves the same over stdio |
546
+ | `/api/mcp/oauth/register` | POST | an AI app registers itself (RFC 7591), before anyone signs it in. Open, as the protocol needs — no key. `{ redirect_uris, client_name?, token_endpoint_auth_method?, grant_types?, response_types?, client_uri? }`: one to ten redirect URIs, each `https` or `http` to localhost, no fragment; `token_endpoint_auth_method` `none` (PKCE alone), `client_secret_post` or `client_secret_basic` (the default). Answers `201 { client_id, client_id_issued_at, client_name, redirect_uris, grant_types, response_types, token_endpoint_auth_method, client_secret?, client_secret_expires_at? }` — the secret only for the two secret methods, shown once. Errors are OAuth's `{ error, error_description }` (`invalid_redirect_uri`, `invalid_client_metadata`); `503` past 5,000 apps on the install |
547
+ | `/api/mcp/oauth/authorize` | GET, POST | the person's half of the sign-in, for the page at `/mcp/authorize` — a signed-in person only, never a key or a support view. GET `?request=<the app's query>` checks it and answers `{ client: { name, id, kind, uri }, redirectHost, loopback, scopes, account, person: { email, role }, workspaces, allWorkspaces }`; POST `{ request, decision: approve\|deny, access?: read\|write, workspaces?: [names] \| null }` answers `{ ok, redirect }` — the app's redirect URI with a one-time `code` (ten minutes), `state` and `iss`, or with `error=access_denied`. A request that names an unknown app or a redirect URI it did not register is `400` and never redirected; the rest of its faults (no PKCE `S256` challenge, a `resource` that is not this `/api/mcp`, an unknown scope) come back as `{ redirect }` with OAuth's error. `access: write` needs the app to have asked for `mcp:write`; `workspaces` must be ones the person may open (`403` otherwise) |
548
+ | `/api/mcp/oauth/token` | POST | OAuth 2.1's token endpoint, form-encoded (JSON is read too), no key: `grant_type=authorization_code` with `code`, `redirect_uri`, `code_verifier` (PKCE S256) and `client_id` — or the app's secret, Basic or in the form —, or `grant_type=refresh_token` with `refresh_token`. `resource`, when sent, must be this `/api/mcp`. Answers `{ access_token, token_type: "Bearer", expires_in: 3600, refresh_token, scope }` with `Cache-Control: no-store`. A code is good once, right or wrong. A refresh token is rotated on use; one used twice revokes the whole connection. Errors: `400 invalid_grant` / `invalid_target` / `unsupported_grant_type`, `401 invalid_client` |
549
+ | `/api/mcp/connections` | GET | the AI apps connected over OAuth: `{ connections: [{ id, client, clientId, redirectHost, person, role, workspaces, access: read\|write, createdAt, lastUsedAt, revokedAt }], everyone }` — every one for `keys:manage` (`everyone: true`), else your own. `?all=1` adds the revoked |
550
+ | `/api/mcp/connections/<id>` | DELETE | disconnect one: the grant is revoked and every token it issued stops at once. Your own, or anyone's with `keys:manage` (`403` otherwise); `404` when there is none or it is already revoked. Audited as `mcp.revoked` |
551
+
552
+ **OAuth for AI apps.** claude.ai — and any MCP client that signs in rather
553
+ than taking a pasted key — finds this from the `401` of `/api/mcp`:
554
+ `WWW-Authenticate: Bearer resource_metadata="…/.well-known/oauth-protected-resource/api/mcp"`
555
+ (RFC 9728), which names this install as the authorization server, whose
556
+ `/.well-known/oauth-authorization-server` (RFC 8414) lists the endpoints above,
557
+ `code_challenge_methods_supported: ["S256"]` and
558
+ `client_id_metadata_document_supported: true` (an app may name itself by an
559
+ `https` URL to its metadata document, fetched through the SSRF guard). Scopes
560
+ are `mcp:read` and `mcp:write`. The access token (`fmcp_at_…`, an hour) opens
561
+ `POST /api/mcp` and nothing else — sent to any other route it is `403` — and
562
+ acts as the person who approved it, at the lower of the role they chose then
563
+ and the one they hold now, in the workspaces both allow; a read-only
564
+ connection is refused every write.
package/docs/cli.md CHANGED
@@ -96,6 +96,8 @@ claude.ai or Claude Code there, nor any MCP configuration on disk.
96
96
  | `foldrun inbox` | what waits on you across your workspaces ([your inbox](runs#your-inbox)): gates, questions agents ask, open issues, proposed fixes, your runs that went wrong this week — assigned to you and started by you by default; `--team`, `--started`, `--mine`, `--all`, `--dismissed`, `--kind approval\|question\|issue\|proposal\|run`, `--unread`, `--json`. Each item shows its id; exits 1 when something assigned to you is unread |
97
97
  | `foldrun inbox read\|unread\|dismiss\|restore <id>…` | mark items in your own view: `dismiss` sets one aside (it still waits where it lives), `restore` puts it back |
98
98
  | `foldrun mcp` | this account as an [MCP server](mcp) on stdin/stdout, for an AI app to start: `claude mcp add foldrun -- foldrun mcp`, `codex mcp add foldrun -- foldrun mcp`. Each line is one JSON-RPC message, forwarded to the platform's `POST /api/mcp` with this machine's sign-in (`--profile`, `--url`, `--token` as for any platform command); stdout carries nothing but protocol. `--read-only` offers only the tools that read. Typed into a terminal it prints how to connect and exits 2 |
99
+ | `foldrun mcp connections` | the AI apps connected to the account over OAuth (claude.ai and the like): the app, where its answers go, whose approval, read only or read and act, which workspaces, connected and last used. Everyone's for an admin (`keys:manage`), else your own; `--all` adds the disconnected, `--json` the raw list |
100
+ | `foldrun mcp disconnect <id>` | end one connection: every token it holds stops at once. Your own, or anyone's as an admin. Asks first; `--yes` when there is no terminal |
99
101
  | `foldrun issues` | the account's [issues](runs#issues): what went wrong, grouped, open until someone resolves it — kind, desk and flow (or the platform cause and every desk it hit), how many times since when, the newest reason and what to do. `--to <workspace>` narrows it, `--all` adds resolved ones (`--status open\|acknowledged\|snoozed\|resolved\|all`), `--json` |
100
102
  | `foldrun issues show\|ack\|snooze\|resolve\|reopen <id>` | one issue: `show` adds its runs and history; `ack` says you have seen it, `snooze --for 30m\|4h\|1d\|1w` (or `--until <ISO>`, or `--times <n>`: until it happens n more times) hides it for a while, `resolve` closes it until it happens again, `reopen` undoes that. `--note "…"` is kept in its history |
101
103
  | `foldrun issues doctor <id>` | the run doctor on an issue's newest run, in that run's workspace — the issue card's **Doctor** button: why (platform, check, work or outside), and the fix as diffs, saved on yes (`--yes`, `--dry-run`, `--json`). A missed schedule has no run and is refused |
@@ -231,7 +233,7 @@ claude.ai or Claude Code there, nor any MCP configuration on disk.
231
233
  | `--payload <json\|text\|@file>` | `approve`: release a step waiting on `wait: event` with this body (`POST …/approve` `{ payload }`). JSON is sent as JSON, anything else as text; `@file` reads a file |
232
234
  | `--note "<text>"` | `approve`: guidance the approved step reads in its prompt — "approve, but skip the Sydney batch". `reject`: the reason, recorded on the trace |
233
235
  | `--read-only` | `mcp`: offer the AI app only the tools that read — no `run_flow`, `decide_approval`, `answer_question`, `message_run` or `stop_run` |
234
- | `--yes` | skip the confirmation, deliberately: `approve`, `stop`, `deploy` (when it would delete files), and a bulk `stop` or `rerun`, `flow rotate-hook`, `flow draft`, `assistant`, `a2a cancel`, `history restore`, `restore`, `billing plan`, `repo deploy`/`merge`, `workspace set name`, and every delete — `runs rm`, `storage rm`, `storage unshare`, `source rm`, `secrets rm`, `keys revoke`, `workspaces rm`. Required when there is no terminal to ask; without it nothing is deleted |
236
+ | `--yes` | skip the confirmation, deliberately: `approve`, `stop`, `deploy` (when it would delete files), and a bulk `stop` or `rerun`, `flow rotate-hook`, `flow draft`, `assistant`, `mcp disconnect`, `a2a cancel`, `history restore`, `restore`, `billing plan`, `repo deploy`/`merge`, `workspace set name`, and every delete — `runs rm`, `storage rm`, `storage unshare`, `source rm`, `secrets rm`, `keys revoke`, `workspaces rm`. Required when there is no terminal to ask; without it nothing is deleted |
235
237
  | `--json` | `report`: the raw run record instead of the report. `observe`, `usage`, `workspaces`, `workspaces show`, `agent ls`, `flow ls`, `agent import --list`, `assistant`, `backups requests`, `status --platform --history`, `team`: the raw document |
236
238
  | `--events <a,b>` | `account set notify`: which events are notified on — `failed`, `awaiting-approval`, `completed` |
237
239
  | `--watch` | `invoke`: follow the run's trace here as it happens; the exit code is the run's |
package/docs/mcp.md CHANGED
@@ -9,16 +9,41 @@ the decisions — approve, answer, message, start, stop.
9
9
  Ask it "what is waiting on me in foldrun?", "why did last night's run of
10
10
  `weekly` fail?" or "approve the draft in run r-…, and say why".
11
11
 
12
- There are two ways in, and they serve the same tools:
12
+ There are three ways in, and they serve the same tools:
13
13
 
14
14
  | | needs | for |
15
15
  |---|---|---|
16
+ | OAuth | nothing: you approve the app in foldrun | claude.ai (web and phone), and any app that signs in by itself |
16
17
  | `foldrun mcp` | the CLI, signed in (`foldrun login`) | an app on the same machine — nothing to paste |
17
- | `POST /api/mcp` | an API key | an app that connects over HTTP, or a machine without the CLI |
18
+ | `POST /api/mcp` with a key | an API key | an app that takes a header, or a machine without the CLI |
18
19
 
19
20
  ## Connecting
20
21
 
21
- With the CLI, on a machine that is signed in:
22
+ ### claude.ai, and apps that sign in
23
+
24
+ In claude.ai: **Settings → Connectors → Add custom connector**, name it
25
+ foldrun, and give it your platform's MCP address:
26
+
27
+ ```
28
+ https://<your platform>/api/mcp
29
+ ```
30
+
31
+ Leave the OAuth client ID and secret empty — claude.ai registers itself. When
32
+ you click **Connect**, it sends you to foldrun: sign in if you are not, and a
33
+ page says which app is asking, where the answer goes (`claude.ai`), and lets
34
+ you choose:
35
+
36
+ - **what it may do** — read only (inbox, runs, flows, files), or read and act
37
+ (also start and stop runs, approve gates, answer and message agents);
38
+ - **which workspaces** — every one you may open, or the ones you tick.
39
+
40
+ **Connect** sends you back to claude.ai, connected; **Deny** sends it a no.
41
+ It is the same in the Claude apps on a phone, which share claude.ai's
42
+ connectors, and in any MCP client that follows the MCP authorization spec.
43
+
44
+ ### With the CLI
45
+
46
+ On a machine that is signed in:
22
47
 
23
48
  ```sh
24
49
  claude mcp add foldrun -- foldrun mcp # Claude Code
@@ -34,9 +59,13 @@ Claude Desktop, Cursor and other apps that take a JSON config:
34
59
  `foldrun mcp` forwards each message to the platform the CLI is signed in to,
35
60
  as the account `foldrun whoami` names — `--profile <name>` picks another
36
61
  signed-in account, `--url` and `--token` (or `FOLDRUN_URL`, `FOLDRUN_TOKEN`)
37
- another platform or key. It renews this machine's sign-in once when it starts,
38
- the way every platform command does. Typed into a terminal it prints how to
39
- connect and exits: it is for an app to start.
62
+ another platform or key. The account is pinned when it starts — `foldrun use`
63
+ in another terminal never moves a running session — and the key is read again
64
+ for every message, so a sign-in another command renews is picked up; the
65
+ renewal itself is checked again twice a day. Typed into a terminal it prints
66
+ how to connect and exits: it is for an app to start.
67
+
68
+ ### With a key
40
69
 
41
70
  Over HTTP, with a key from Settings → API keys:
42
71
 
@@ -46,8 +75,30 @@ claude mcp add --transport http foldrun https://<your platform>/api/mcp \
46
75
  ```
47
76
 
48
77
  Settings → API keys → **Connecting an AI app (MCP)** shows this install's
49
- address and the same lines. claude.ai's connectors sign in with OAuth, which
50
- this endpoint does not offer; connect through Claude Desktop or Claude Code.
78
+ address and the same lines.
79
+
80
+ ## Connected apps
81
+
82
+ Every app connected over OAuth is listed in Settings → API keys → **Connected
83
+ AI apps**: the app, where its answers went, whose approval, what access and
84
+ which workspaces, when it was connected and last used. **Disconnect** ends it —
85
+ every token it holds stops at once — and the person can connect it again. You
86
+ see your own; an admin sees everyone's and may disconnect any.
87
+
88
+ ```sh
89
+ foldrun mcp connections # yours, or everyone's for an admin (--all adds the disconnected)
90
+ foldrun mcp disconnect <id> # asks first; --yes on a pipe
91
+ ```
92
+
93
+ What a connection may do is decided at every call, not only when it was
94
+ approved: it acts as the person who approved it, at the lower of the role
95
+ they chose then and the role they hold now, in the workspaces both allow. A
96
+ person who leaves the account, or is made a viewer, takes their apps with
97
+ them. The access token lasts an hour and opens `/api/mcp` only — sent to any
98
+ other route it is refused. Refresh tokens rotate on every use, and one used
99
+ twice disconnects the app: someone else has a copy. Approving, disconnecting
100
+ and that automatic disconnect are in the audit log (`mcp.connected`,
101
+ `mcp.revoked`).
51
102
 
52
103
  ## The tools
53
104
 
@@ -115,11 +166,22 @@ answered `202` with no body. It sits under `/api`, so the proxy holds the same
115
166
  raw-path check, body limit and rate limit as for every other route.
116
167
 
117
168
  Refused before a message is read, as a JSON-RPC error with a real status: no
118
- key (`401`, with `WWW-Authenticate: Bearer`) — a session cookie never opens
119
- it —, a refused key (`401`/`403`), an `Origin` header that is not the
120
- platform's own (`403`; MCP clients send none), an `MCP-Protocol-Version` it
121
- does not speak (`400`). The route and request shape are in [the API
122
- reference](api#mcp--ai-apps-connecting-with-a-key).
169
+ credential (`401`, with `WWW-Authenticate: Bearer resource_metadata="…/.well-known/oauth-protected-resource/api/mcp"`,
170
+ which is where an app that signs in starts) — a session cookie never opens
171
+ it —, a refused key or an expired or revoked token (`401`, `error="invalid_token"`),
172
+ an `Origin` header that is not the platform's own (`403`; MCP clients send
173
+ none), an `MCP-Protocol-Version` it does not speak (`400`). The route and
174
+ request shape are in [the API reference](api#mcp--ai-apps-connecting-with-a-key).
175
+
176
+ The sign-in follows the MCP authorization spec on OAuth 2.1:
177
+ `/.well-known/oauth-protected-resource` (RFC 9728) names this install as the
178
+ authorization server; `/.well-known/oauth-authorization-server` (RFC 8414)
179
+ lists the endpoints, PKCE `S256` only, and `mcp:read` / `mcp:write`; an app
180
+ registers itself (RFC 7591) or names itself by an `https` URL to its own
181
+ metadata document, which foldrun fetches through its SSRF guard; every token is
182
+ bound to this `/api/mcp` (RFC 8707). Redirect URIs are `https`, or `http` to
183
+ localhost for a desktop app — and for those the approval page warns that the
184
+ app is on this computer.
123
185
 
124
186
  `foldrun mcp` reads one message per line on stdin and writes one per line on
125
187
  stdout, and nothing else there: every note goes to stderr. Calls run side by
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "foldrun",
3
- "version": "10.3.1",
3
+ "version": "10.4.0",
4
4
  "description": "Agents are just folders. Write them, check them, run them — on your machine.",
5
5
  "license": "FSL-1.1-ALv2",
6
6
  "type": "module",
package/src/commands.mjs CHANGED
@@ -6632,7 +6632,17 @@ function issueDoctor(issue) {
6632
6632
  * @param {string[]} positional @param {Record<string, any>} flags
6633
6633
  */
6634
6634
  async function mcpCmd(positional, flags) {
6635
- if (positional.length) throw new Error("`foldrun mcp` takes no arguments — an AI app starts it: `claude mcp add foldrun -- foldrun mcp` (--read-only for the tools that read)");
6635
+ const [verb, arg, ...extra] = positional;
6636
+ if (verb === "connections" || verb === "ls") return mcpConnectionsCmd(flags);
6637
+ if (verb === "disconnect" || verb === "revoke") {
6638
+ if (!arg || extra.length) throw new Error("which connection? `foldrun mcp disconnect <id>` — ids are in `foldrun mcp connections`");
6639
+ const url = platformFor(flags, "mcp disconnect");
6640
+ if (!(await sureToDelete(flags, "mcp disconnect", `the AI app connection ${arg} on ${url} (every token it holds stops working)`))) return 1;
6641
+ await remoteCall(url, flags, `/api/mcp/connections/${encodeURIComponent(arg)}`, { method: "DELETE" });
6642
+ console.log(`\n ${c.green("✓")} disconnected ${arg}\n`);
6643
+ return 0;
6644
+ }
6645
+ if (positional.length) throw new Error("`foldrun mcp` serves the tools to an AI app (`claude mcp add foldrun -- foldrun mcp`); its verbs are connections and disconnect <id> — not " + verb);
6636
6646
  if (process.stdin.isTTY && flags.yes !== true) {
6637
6647
  console.error(`\n ${c.bold("foldrun mcp")} is an MCP server on stdin/stdout, for an AI app to start — not to type into.\n\n claude mcp add foldrun -- foldrun mcp${flags["read-only"] === true ? " --read-only" : ""}\n codex mcp add foldrun -- foldrun mcp\n\n ${c.dim("Claude Desktop, Cursor: { \"mcpServers\": { \"foldrun\": { \"command\": \"foldrun\", \"args\": [\"mcp\"] } } } — `foldrun docs mcp`")}\n`);
6638
6648
  return 2;
@@ -6737,6 +6747,28 @@ async function mcpCmd(positional, flags) {
6737
6747
  return 0;
6738
6748
  }
6739
6749
 
6750
+ /** `foldrun mcp connections` — the AI apps connected to the account over OAuth
6751
+ * (claude.ai and the like): which app, whose approval, what access. Every
6752
+ * one for an admin; your own otherwise. */
6753
+ async function mcpConnectionsCmd(flags) {
6754
+ const url = platformFor(flags, "mcp connections");
6755
+ const view = await remoteCall(url, flags, `/api/mcp/connections${flags.all === true ? "?all=1" : ""}`);
6756
+ if (flags.json) {
6757
+ console.log(JSON.stringify(view, null, 2));
6758
+ return 0;
6759
+ }
6760
+ const rows = view.connections ?? [];
6761
+ console.log(`\n ${c.bold("AI apps connected over OAuth")} ${c.dim(view.everyone ? "· everyone's" : "· yours")}`);
6762
+ if (!rows.length) console.log(`\n ${c.dim("none — claude.ai connects itself when you add this platform's /api/mcp as a custom connector")}`);
6763
+ for (const r of rows) {
6764
+ const used = r.lastUsedAt ? `last used ${String(r.lastUsedAt).slice(0, 16).replace("T", " ")}` : "not used yet";
6765
+ console.log(`\n ${c.bold(r.client)} ${c.dim(`· ${r.redirectHost} · ${r.access === "read" ? "read only" : "read and act"} · ${r.workspaces ? r.workspaces.join(", ") : "all workspaces"}`)}${r.revokedAt ? c.red(" · disconnected") : ""}`);
6766
+ console.log(` ${c.dim(`${r.id} · ${r.person} · connected ${String(r.createdAt).slice(0, 16).replace("T", " ")} · ${used}`)}`);
6767
+ }
6768
+ console.log("");
6769
+ return 0;
6770
+ }
6771
+
6740
6772
  async function inboxCmd(positional, flags) {
6741
6773
  const url = platformFor(flags, "inbox");
6742
6774
  const [verb, ...ids] = positional;