@zackbart/connecta 0.28.1 → 0.29.0

package/dist/providers/notion/skill.generated.js

@@ -0,0 +1,10 @@
+// Generated by scripts/generate-providers.mjs; do not edit.
+export const skill = {
+    "content": "\n## Identity before routing\n\nThis explicit `surface: \"api\"` connection acts as an internal integration.\nIts `integration_*` tools are canonical for bot-scoped/headless operations.\nHosted MCP is the default and acts as its OAuth-authorized user. Authorization\nis not equivalent: an integration token cannot authorize hosted MCP, and user\nOAuth cannot reproduce the integration bot's ownership/sharing boundary.\nConfigure each under a separate connector id and authorize it independently.\nDo not switch to a user identity to recover a failed integration operation.\n\nUse the bot tools when the task requires that integration identity. Use hosted\n`notion-search`, `notion-fetch`, `notion-get-users`, `notion-get-comments`,\n`notion-create-pages`, `notion-update-page`, and `notion-create-comment` for\nuser-owned content work. Hosted MCP additionally covers sessions, agents,\nattachments and broader workspace operations. REST tool names are not MCP aliases.\n\n\n\n## Databases contain data sources\n\nA Notion database is a container; the rows and the schema live in a *data\nsource* inside it. The two ids are not interchangeable and Notion rejects the\nwrong one.\n\n- The id in a Notion database URL is a **database id**.\n- `integration_get_database` turns it into the `data_sources` list — usually one entry.\n- `integration_get_data_source_schema` and `integration_query_data_source` take that\n  **data_source_id**, and so does `integration_create_page` when adding a row.\n\nSo the sequence for \"find rows in this database\" is `integration_get_database` →\n`integration_get_data_source_schema` → `integration_query_data_source`. `integration_search` returns data\nsources directly and skips the first step.\n\n## Property quirks that break writes\n\n- Property names in filters, sorts, and writes must match the schema exactly,\n  including case. Read `integration_get_data_source_schema` before composing one.\n- `select` and `status` writes must use an existing option name; inventing\n  one fails. The schema lists the valid options.\n- Writes **replace** a property. Sending one item to a `multi_select` or\n  `relation` drops the others, so send the complete intended value.\n- `rollup`, `formula`, `unique_id`, and the created/edited fields are\n  computed. They cannot be written.\n- A page's title column is rarely called \"title\" — pass `title_property` from\n  the schema when creating or updating a row.\n- Notion truncates `title`, `rich_text`, `relation`, and `people` at 25\n  entries. `integration_get_page` reports each one in `truncated_properties` as\n  `{ name, id }`; pass that `id` as `property_id` to\n  `integration_get_page_property` for the complete value.\n\n## Reading page content\n\n`integration_get_page` returns properties. `integration_get_page_content` returns the body as flat\nblocks. Nested blocks (toggles, list children, table rows) need `depth`, and\neach level multiplies requests — a deep read stops at an internal ceiling and\nreports `truncated: true` rather than spending the whole rate-limit budget.\n\n## Appending is append-only\n\n`integration_append_blocks` adds children and nothing else. It cannot move, reorder, or\nreplace an existing block, and a block appended through the API can never be\nrelocated by it afterwards. Get the position right the first time with\n`position` and `after_block_id`. Notion caps one call at 100 blocks.\n\n## Lean by default, raw on request\n\nEvery read projects Notion's payload down to ids, plain text, and flattened\nproperty values. Where the dropped detail can matter — `integration_search`, `integration_get_page`,\n`integration_get_page_content`, `integration_get_page_property`, `integration_get_data_source_schema`,\n`integration_query_data_source`, `integration_list_comments` — pass `raw: true` to get Notion's\nuntouched response instead. It is much larger; reach for it only when a\nspecific field is missing. Narrow `integration_query_data_source` and `integration_get_page` with\n`properties` instead whenever the goal is fewer fields, not more.\n\n`integration_get_page_content` with `raw: true` returns one level exactly as Notion\nsent it and does not walk nested children, so `depth` is ignored alongside\nit. A block type this projection does not model keeps its payload under\n`raw` on the block itself, so nothing silently flattens to an empty string.\n\n## Failures worth reading carefully\n\n- **404** means \"no such object\" *or* \"not shared with this integration\", and\n  Notion will not say which. Never conclude a page was deleted from it; check\n  that the page is shared with the integration in Notion.\n- **403** is not an expired token. The integration is missing a capability\n  (comment capabilities are off by default) or the object was never shared.\n  Re-authorizing cannot fix it; an operator must change it in Notion.\n- **429** carries a retry window. Notion allows roughly three requests per\n  second per integration, so wait it out rather than retrying immediately.\n\n## No escape hatch\n\nThis connection has no guarded raw-REST tool, deliberately. Notion's public\nAPI is small and slow-moving enough for the named surface to cover it, so\nthere is no `notion_api_*` to reach for — an operation absent from the tool\nlist is absent from this connection, not hidden behind a generic call.\n\n## Writes and pagination\n\n- Notion has **no idempotency key**. A retried `integration_create_page` or\n  `integration_add_comment` creates a duplicate. Confirm with `integration_search` before repeating\n  a write that may have partially succeeded.\n- List tools take `page_size` (max 100) and return `has_more` with\n  `next_cursor`. Pass a cursor back verbatim — it is opaque and must never be\n  parsed or constructed. Follow pages inside `execute_code` and reduce there.\n# Notion MCP usage\n\nOfficial MCP interface: tool names, descriptions, argument schemas, and result\nschemas come from Notion's live server. Connecta preserves that catalog and\napplies release-reviewed classification. Reviewed writes stay writes even when Notion claims they only read.\n\nWorkspace purpose: \n\n- This connection acts as its OAuth-authorized user. The separately configured\n  `surface: \"api\"` connection retains required internal-integration operations\n  under `integration_*` names, including headless content writes, cursor-complete\n  queries, exact JSON blocks and property items. No identity fallback occurs.\n- Discover the live catalog before assuming a tool exists. Notion can gate\n  tools by workspace, account, client, and rollout independently of Connecta.\n- Start broad discovery with `notion-search`, then use `notion-fetch` on\n  the exact page, database, data source, or object before changing it.\n- Use the live input schema as the contract. Notion owns these MCP schemas;\n  the REST schemas in Connecta's API interface do not apply to MCP tools with\n  similar names.\n- Session and agent tools can launch asynchronous work. Read session state and\n  events before sending another message, waiting, or stopping a session.\n- File uploads and attachment tools create durable workspace state. Keep\n  downloads inside the requested task and do not expose signed attachment URLs.\n- An `auth_required` failure means this connector's OAuth grant is missing or\n  expired. Run `authorize_connector` for this connector id, then retry.\n",
+    "instructionsHeading": "Workspace instructions",
+    "fragments": {
+        "guide_0": "\n## Identity before routing\n\nThis explicit `surface: \"api\"` connection acts as an internal integration.\nIts `integration_*` tools are canonical for bot-scoped/headless operations.\nHosted MCP is the default and acts as its OAuth-authorized user. Authorization\nis not equivalent: an integration token cannot authorize hosted MCP, and user\nOAuth cannot reproduce the integration bot's ownership/sharing boundary.\nConfigure each under a separate connector id and authorize it independently.\nDo not switch to a user identity to recover a failed integration operation.\n\nUse the bot tools when the task requires that integration identity. Use hosted\n`notion-search`, `notion-fetch`, `notion-get-users`, `notion-get-comments`,\n`notion-create-pages`, `notion-update-page`, and `notion-create-comment` for\nuser-owned content work. Hosted MCP additionally covers sessions, agents,\nattachments and broader workspace operations. REST tool names are not MCP aliases.\n\n\n\n## Databases contain data sources\n\nA Notion database is a container; the rows and the schema live in a *data\nsource* inside it. The two ids are not interchangeable and Notion rejects the\nwrong one.\n\n- The id in a Notion database URL is a **database id**.\n- `integration_get_database` turns it into the `data_sources` list — usually one entry.\n- `integration_get_data_source_schema` and `integration_query_data_source` take that\n  **data_source_id**, and so does `integration_create_page` when adding a row.\n\nSo the sequence for \"find rows in this database\" is `integration_get_database` →\n`integration_get_data_source_schema` → `integration_query_data_source`. `integration_search` returns data\nsources directly and skips the first step.\n\n## Property quirks that break writes\n\n- Property names in filters, sorts, and writes must match the schema exactly,\n  including case. Read `integration_get_data_source_schema` before composing one.\n- `select` and `status` writes must use an existing option name; inventing\n  one fails. The schema lists the valid options.\n- Writes **replace** a property. Sending one item to a `multi_select` or\n  `relation` drops the others, so send the complete intended value.\n- `rollup`, `formula`, `unique_id`, and the created/edited fields are\n  computed. They cannot be written.\n- A page's title column is rarely called \"title\" — pass `title_property` from\n  the schema when creating or updating a row.\n- Notion truncates `title`, `rich_text`, `relation`, and `people` at 25\n  entries. `integration_get_page` reports each one in `truncated_properties` as\n  `{ name, id }`; pass that `id` as `property_id` to\n  `integration_get_page_property` for the complete value.\n\n## Reading page content\n\n`integration_get_page` returns properties. `integration_get_page_content` returns the body as flat\nblocks. Nested blocks (toggles, list children, table rows) need `depth`, and\neach level multiplies requests — a deep read stops at an internal ceiling and\nreports `truncated: true` rather than spending the whole rate-limit budget.\n\n## Appending is append-only\n\n`integration_append_blocks` adds children and nothing else. It cannot move, reorder, or\nreplace an existing block, and a block appended through the API can never be\nrelocated by it afterwards. Get the position right the first time with\n`position` and `after_block_id`. Notion caps one call at 100 blocks.\n\n## Lean by default, raw on request\n\nEvery read projects Notion's payload down to ids, plain text, and flattened\nproperty values. Where the dropped detail can matter — `integration_search`, `integration_get_page`,\n`integration_get_page_content`, `integration_get_page_property`, `integration_get_data_source_schema`,\n`integration_query_data_source`, `integration_list_comments` — pass `raw: true` to get Notion's\nuntouched response instead. It is much larger; reach for it only when a\nspecific field is missing. Narrow `integration_query_data_source` and `integration_get_page` with\n`properties` instead whenever the goal is fewer fields, not more.\n\n`integration_get_page_content` with `raw: true` returns one level exactly as Notion\nsent it and does not walk nested children, so `depth` is ignored alongside\nit. A block type this projection does not model keeps its payload under\n`raw` on the block itself, so nothing silently flattens to an empty string.\n\n## Failures worth reading carefully\n\n- **404** means \"no such object\" *or* \"not shared with this integration\", and\n  Notion will not say which. Never conclude a page was deleted from it; check\n  that the page is shared with the integration in Notion.\n- **403** is not an expired token. The integration is missing a capability\n  (comment capabilities are off by default) or the object was never shared.\n  Re-authorizing cannot fix it; an operator must change it in Notion.\n- **429** carries a retry window. Notion allows roughly three requests per\n  second per integration, so wait it out rather than retrying immediately.\n\n## No escape hatch\n\nThis connection has no guarded raw-REST tool, deliberately. Notion's public\nAPI is small and slow-moving enough for the named surface to cover it, so\nthere is no `notion_api_*` to reach for — an operation absent from the tool\nlist is absent from this connection, not hidden behind a generic call.\n\n## Writes and pagination\n\n- Notion has **no idempotency key**. A retried `integration_create_page` or\n  `integration_add_comment` creates a duplicate. Confirm with `integration_search` before repeating\n  a write that may have partially succeeded.\n- List tools take `page_size` (max 100) and return `has_more` with\n  `next_cursor`. Pass a cursor back verbatim — it is opaque and must never be\n  parsed or constructed. Follow pages inside `execute_code` and reduce there.\n",
+        "guide_1": "# Notion MCP usage\n\nOfficial MCP interface: tool names, descriptions, argument schemas, and result\nschemas come from Notion's live server. Connecta preserves that catalog and\napplies release-reviewed classification. Reviewed writes stay writes even when Notion claims they only read.\n\nWorkspace purpose: ",
+        "guide_2": "\n\n- This connection acts as its OAuth-authorized user. The separately configured\n  `surface: \"api\"` connection retains required internal-integration operations\n  under `integration_*` names, including headless content writes, cursor-complete\n  queries, exact JSON blocks and property items. No identity fallback occurs.\n- Discover the live catalog before assuming a tool exists. Notion can gate\n  tools by workspace, account, client, and rollout independently of Connecta.\n- Start broad discovery with `notion-search`, then use `notion-fetch` on\n  the exact page, database, data source, or object before changing it.\n- Use the live input schema as the contract. Notion owns these MCP schemas;\n  the REST schemas in Connecta's API interface do not apply to MCP tools with\n  similar names.\n- Session and agent tools can launch asynchronous work. Read session state and\n  events before sending another message, waiting, or stopping a session.\n- File uploads and attachment tools create durable workspace state. Keep\n  downloads inside the requested task and do not expose signed attachment URLs.\n- An `auth_required` failure means this connector's OAuth grant is missing or\n  expired. Run `authorize_connector` for this connector id, then retry.\n"
+    }
+};

package/dist/providers/{overflow.d.ts → overflow/index.d.ts}

@@ -1,4 +1,4 @@
-import type { Connector, ConnectorCallAdmissionPolicy } from "../types.js";
+import type { ConnectorCallAdmissionPolicy } from "../../types.js";
 /**
  * Overflow's two published environments. They are separate deployments with
  * separate credentials: a staging key works only against staging and a
@@ -57,4 +57,4 @@ export interface OverflowOptions {
  * cannot infer — Overflow's keys carry no recognizable environment marker — so
  * the operator states it, and the title and the guide's first line repeat it.
  */
-export declare function overflow(id: string, options: OverflowOptions): Connector;
+export declare const overflow: import("../../provider.js").ProviderFactory<OverflowOptions>;

package/dist/providers/{overflow.js → overflow/index.js}

@@ -1,3 +1,4 @@
+import { skill } from "./skill.generated.js";
 /**
  * Overflow (overflow.co) — church and nonprofit giving: cash, stock, crypto,
  * and donor-advised-fund gifts, the deposits that settle them, the donors who
@@ -7,7 +8,7 @@
  * header-authenticated JSON over thirty-eight operations, so `fetch` through
  * the shared guarded transport keeps this subpath Workers-clean and adds
  * nothing to install. Every tool below is hand-written. Overflow's OpenAPI
- * document is drift evidence only — `scripts/drift/overflow-endpoints.json`
+ * document is drift evidence only — `src/providers/overflow/drift.json`
  * records the operations the named tools touch, and
  * `npm run providers:check -- --provider overflow` compares them with the
  * published document without a credential. That document is served only by
@@ -27,9 +28,12 @@
  * exact method, path, and body a human is being asked to allow. Overflow has
  * no upload endpoint, so there is no upload hatch.
  */
-import { apiConnector as api, defined } from "../connectors/api-connector.js";
-import { guardedFetch, retryAfterMs, } from "../connectors/guarded-fetch.js";
-import { ConnectorCallError } from "../errors.js";
+import { apiConnector as api, defined } from "../../connectors/api-connector.js";
+import { guardedFetch, retryAfterMs, } from "../../connectors/guarded-fetch.js";
+import { ConnectorCallError } from "../../errors.js";
+import { keys, optionsOf } from "../../config-schema.js";
+import { PROVIDER_COMMON } from "../../connectors/option-shapes.js";
+import { asProviderFactory } from "../../provider.js";
 /**
  * Overflow's two published environments. They are separate deployments with
  * separate credentials: a staging key works only against staging and a
@@ -81,9 +85,7 @@ const OVERFLOW_ADMISSION = {
     ],
 };
 function asRecord(value) {
-    return value && typeof value === "object" && !Array.isArray(value)
-        ? value
-        : {};
+    return value && typeof value === "object" && !Array.isArray(value) ? value : {};
 }
 function asArray(value) {
     return Array.isArray(value) ? value : [];
@@ -94,9 +96,7 @@ function compact(value) {
 /** A `{ id, name }` reference, or undefined when Overflow sent none. */
 function reference(value) {
     const record = asRecord(value);
-    return Object.keys(record).length === 0
-        ? undefined
-        : compact({ id: record["id"], name: record["name"] });
+    return Object.keys(record).length === 0 ? undefined : compact({ id: record["id"], name: record["name"] });
 }
 // --- Credentials and transport ----------------------------------------------
 const API_KEY_CREDENTIAL = {
@@ -176,9 +176,7 @@ function windowResetMs(headers) {
     if (!raw)
         return undefined;
     const seconds = Number(raw.trim());
-    return Number.isFinite(seconds) && seconds >= 0
-        ? Math.trunc(seconds * 1_000)
-        : undefined;
+    return Number.isFinite(seconds) && seconds >= 0 ? Math.trunc(seconds * 1_000) : undefined;
 }
 /**
  * Map an Overflow failure by what the caller does next (H11).
@@ -203,7 +201,7 @@ function overflowFailure(status, headers, payload, method) {
         const ambiguity = method !== "GET" && status === 403
             ? " Overflow also answers 403 when a donor profile owned by a donor account cannot be edited through the API; if reads succeed with this credential, that refusal is the cause and no credential change will fix it."
             : "";
-        return new ConnectorCallError("auth_required", `Overflow rejected the request (HTTP ${status}): ${detail} The client id and API key may be wrong, revoked, or issued for the other environment (staging and production credentials are not interchangeable); an operator must replace them.${ambiguity}`);
+        return new ConnectorCallError(status === 403 ? "provider_permission_denied" : "auth_required", `Overflow rejected the request (HTTP ${status}): ${detail} The client id and API key may be wrong, revoked, or issued for the other environment (staging and production credentials are not interchangeable); an operator must check them and ask the account administrator to grant resource access if they are valid.${ambiguity}`);
     }
     if (status === 404) {
         return new ConnectorCallError("not_found", `Overflow found no such resource for this nonprofit (HTTP 404): ${detail} Re-resolve the id with its list tool; for a raw call, confirm the path below /api/v3.`);
@@ -221,7 +219,9 @@ function overflowFailure(status, headers, payload, method) {
         const wait = windowResetMs(headers);
         return new ConnectorCallError("unavailable", `Overflow is failing upstream (HTTP ${status}): ${detail}`, wait === undefined ? {} : { retryAfterMs: wait });
     }
-    return new ConnectorCallError("connector_call_failed", `Overflow request failed (HTTP ${status}): ${detail}`, { retryable: false });
+    return new ConnectorCallError("connector_call_failed", `Overflow request failed (HTTP ${status}): ${detail}`, {
+        retryable: false,
+    });
 }
 function parseBody(text) {
     if (!text.trim())
@@ -311,9 +311,7 @@ function projectContribution(value) {
         stocks: Object.keys(stocks).length === 0
             ? undefined
             : compact({ quantity: stocks["quantity"], tickers: stocks["tickers"] }),
-        crypto: Object.keys(crypto).length === 0
-            ? undefined
-            : compact({ quantity: crypto["quantity"], token: crypto["token"] }),
+        crypto: Object.keys(crypto).length === 0 ? undefined : compact({ quantity: crypto["quantity"], token: crypto["token"] }),
         dedication: contribution["dedication"] ?? undefined,
         donorNotes: contribution["donorNotes"] ?? undefined,
         metadata: contribution["metadata"] ?? undefined,
@@ -329,7 +327,7 @@ function projectDonor(value, full) {
         lastName: donor["lastName"],
         email: donor["email"],
         phone: full ? donor["phone"] : undefined,
-        address: full ? donor["address"] ?? undefined : undefined,
+        address: full ? (donor["address"] ?? undefined) : undefined,
         locationIds: donor["locationIds"],
         totalContributionsCount: donor["totalContributionsCount"],
         activeRecurringCount: donor["activeRecurringCount"],
@@ -444,7 +442,9 @@ function projectPaymentMethod(value) {
 function projectCampaign(value) {
     const campaign = asRecord(value);
     const refs = (list) => {
-        const rows = asArray(list).map(reference).filter((row) => row !== undefined);
+        const rows = asArray(list)
+            .map(reference)
+            .filter((row) => row !== undefined);
         return rows.length === 0 ? undefined : rows;
     };
     return compact({
@@ -840,10 +840,7 @@ function rawQuery(value) {
             (query[name] ??= []).push(item);
         }
     }
-    return Object.fromEntries(Object.entries(query).map(([name, values]) => [
-        name,
-        values.length === 1 ? values[0] : values,
-    ]));
+    return Object.fromEntries(Object.entries(query).map(([name, values]) => [name, values.length === 1 ? values[0] : values]));
 }
 function tools(send, defaultPageSize) {
     const readOnly = { readOnlyHint: true };
@@ -864,7 +861,9 @@ function tools(send, defaultPageSize) {
         const payload = await callOverflow(send, { method: "GET", path }, ctx);
         const record = asRecord(payload)["data"];
         if (record === undefined || record === null) {
-            throw new ConnectorCallError("connector_call_failed", "Overflow answered without the requested record.", { retryable: false });
+            throw new ConnectorCallError("connector_call_failed", "Overflow answered without the requested record.", {
+                retryable: false,
+            });
         }
         return asRecord(record);
     }
@@ -930,7 +929,11 @@ function tools(send, defaultPageSize) {
                     type: "array",
                     minItems: 1,
                     uniqueItems: true,
-                    items: { type: "string", enum: ["PENDING", "CONFIRMED", "FAILED", "CANCELED"], description: "Status bucket." },
+                    items: {
+                        type: "string",
+                        enum: ["PENDING", "CONFIRMED", "FAILED", "CANCELED"],
+                        description: "Status bucket.",
+                    },
                     description: "Overflow's coarse status buckets; matches any.",
                 },
                 campaignId: idProperty("list_campaigns"),
@@ -966,7 +969,9 @@ function tools(send, defaultPageSize) {
             name: "get_contribution",
             description: "Get one contribution by id with its status, amount, campaign, deposit, and donor name. Refunds and chargebacks are separate lists.",
             annotations: readOnly,
-            inputSchema: namedInput({ contributionId: idProperty("list_contributions"), raw: RAW_PROPERTY }, ["contributionId"]),
+            inputSchema: namedInput({ contributionId: idProperty("list_contributions"), raw: RAW_PROPERTY }, [
+                "contributionId",
+            ]),
             outputSchema: CONTRIBUTION_SCHEMA,
             handler: async (args, ctx) => {
                 const record = await one(`/contributions/${segment(args["contributionId"])}`, ctx);
@@ -1189,9 +1194,7 @@ function tools(send, defaultPageSize) {
             outputSchema: listOutput("subscriptions", SUBSCRIPTION_SCHEMA),
             handler: (args, ctx) => list({
                 args,
-                path: args["donorId"] === undefined
-                    ? "/subscriptions"
-                    : `/subscriptions/${segment(args["donorId"])}`,
+                path: args["donorId"] === undefined ? "/subscriptions" : `/subscriptions/${segment(args["donorId"])}`,
                 key: "subscriptions",
                 project: projectSubscription,
                 query: {
@@ -1240,7 +1243,15 @@ function tools(send, defaultPageSize) {
                 excludeInactive: { type: "boolean", description: "Exclude campaigns that have ended." },
                 sortBy: {
                     type: "string",
-                    enum: ["startDate", "endDate", "name", "displayOrder", "totalContributionCount", "totalContributionValue", "uniqueDonorCount"],
+                    enum: [
+                        "startDate",
+                        "endDate",
+                        "name",
+                        "displayOrder",
+                        "totalContributionCount",
+                        "totalContributionValue",
+                        "uniqueDonorCount",
+                    ],
                     description: "Sort field. Defaults to startDate.",
                 },
                 sortDirection: { ...SORT_DIRECTION, description: "Defaults to DESC." },
@@ -1256,7 +1267,9 @@ function tools(send, defaultPageSize) {
                 project: projectCampaign,
                 query: {
                     isSubcampaign: args["isSubcampaign"] === undefined
-                        ? args["parentCampaignId"] === undefined ? undefined : "true"
+                        ? args["parentCampaignId"] === undefined
+                            ? undefined
+                            : "true"
                         : String(args["isSubcampaign"]),
                     parentCampaignId: args["parentCampaignId"],
                     search: args["search"],
@@ -1273,7 +1286,11 @@ function tools(send, defaultPageSize) {
             annotations: readOnly,
             inputSchema: namedInput({
                 includeArchived: { type: "boolean", description: "Include archived locations." },
-                sortBy: { type: "string", enum: ["displayOrder", "name", "createdAt"], description: "Sort field. Defaults to displayOrder." },
+                sortBy: {
+                    type: "string",
+                    enum: ["displayOrder", "name", "createdAt"],
+                    description: "Sort field. Defaults to displayOrder.",
+                },
                 sortDirection: { ...SORT_DIRECTION, description: "Defaults to ASC." },
                 page: PAGE_PROPERTY,
                 limit: LIMIT,
@@ -1395,9 +1412,7 @@ function tools(send, defaultPageSize) {
                 const next = info["nextCursor"];
                 return compact({
                     attempts: args["raw"] === true ? rows : rows.map(projectEventLog),
-                    nextCursor: info["hasNextPage"] === true && typeof next === "string" && next !== ""
-                        ? next
-                        : undefined,
+                    nextCursor: info["hasNextPage"] === true && typeof next === "string" && next !== "" ? next : undefined,
                 });
             },
         },
@@ -1413,78 +1428,14 @@ function usageGuide(purpose, environment, instructions) {
 
 ${environmentLine}
 
-Nonprofit purpose: ${purpose}
-
-## Money, units, and status
-
-Overflow does not use one money unit, so read the unit per field:
-
-- Fields named \`…InCents\` are cents: deposits and their line items and
-  totals, refunds, chargebacks, and the authorize body.
-- Recurring-gift \`amount\` is dollars: Overflow's create and update bodies
-  say so, and its subscription reads carry the same field.
-- Contribution \`amount\` states no unit in Overflow's schema. Check one known
-  gift before reporting totals.
-- \`status\` is per asset type. A gift is final at \`PAID_OUT\` (cash),
-  \`CONFIRMED\` (manual cash, crypto), \`CONTRIBUTION_RECEIVED\` or \`BILLED\`
-  (stock), and \`CONTRIBUTION_RECEIVED\` (DAF). \`statusBucket\` filters by
-  Overflow's coarser PENDING / CONFIRMED / FAILED / CANCELED.
-- There is no funds endpoint. A gift reports its campaign and subcampaign;
-  \`fundId\` and \`subFundId\` appear only in the payment-authorize body.
-- Poll for changes with \`minimumUpdatedDate\`; \`minimumInitiatedDate\` is
-  when the donor gave.
-
-## Reconcile a deposit
-
-\`list_deposits\` → \`get_deposit_summary\` for totals → \`get_deposit\` for the
-line items, whose \`referenceId\` names the contribution, refund, or
-chargeback. Contributions also carry their \`depositId\`.
-
-## Donor data is sensitive
-
-Giving history and donor contact details are personal data. Reduce inside
-\`execute_code\` — aggregate, count, or keep ids — and return names, emails,
-phones, or addresses only when the task needs them. Lists and contributions
-already omit contact details; \`get_donor\` is the deliberate lookup.
-
-## Writes go through \`overflow_api_mutate\`
-
-Every Overflow write moves money or changes a donor record, so none has a
-named tool. Each is one approval-gated \`overflow_api_mutate\` call with
-Overflow's documented body:
-
-- \`POST /contributions\` charges a saved payment method. Its \`amount\`
-  states no unit; confirm it against a known gift of the same donor before
-  charging. \`POST /payments/authorize\` takes \`amountInCents\` and
-  authorizes a new method (\`0\` only saves it).
-- \`POST /contributions/{id}/initiate-refund\` (under $1,000 only).
-- \`POST\` / \`PATCH\` / \`DELETE /subscriptions/{donorId}[/{subscriptionId}]\`
-  creates, changes, or cancels a recurring gift. Its \`amount\` is dollars,
-  not cents: a $50.00 monthly gift is \`amount: 50\`. Cancel takes a
-  \`cancellationReason\` body.
-- \`POST /donors\`, \`PATCH /donors/{id}\`.
-
-Resolve every id first — donors, payment methods (\`list_payment_methods\`),
-campaigns, locations — and never guess one. A write that fails with HTTP 5xx
-may still have happened: read before retrying.
-
-## Hatches, pages, and limits
-
-- \`overflow_api_get\` reaches the reads without a named tool: single refunds,
-  chargebacks, campaigns, locations, webhooks, and recurring gifts
-  (\`/subscriptions/{donorId}/{id}\`), and Tap devices, groups, and
-  destinations (\`/tap/devices\`, \`/tap/groups\`, \`/tap/destinations\`). Paths
-  are below \`/api/v3\`; array filters repeat a \`name[]\` query pair.
-- Lists page by number: pass \`page\` while \`page.hasMore\` is true.
-  \`list_webhook_event_logs\` pages by cursor instead.
-- Overflow allows 120 requests per minute per API client. This connection
-  admits the same per runtime, four at a time, which approximates rather than
-  enforces it; a \`rate_limited\` failure carries the wait.
-${accountInstructions
-        ? `\n## Nonprofit instructions\n\n${accountInstructions}\n`
-        : ""}`;
+Nonprofit purpose: ${purpose}${skill.fragments.guide_0}${accountInstructions ? `\n## ${skill.instructionsHeading}\n\n${accountInstructions}\n` : ""}`;
 }
 // --- Construction --------------------------------------------------------------
+/** The closed options overflow() accepts; see `assertKnownOptions`. */
+const OVERFLOW_OPTIONS = optionsOf()({
+    ...PROVIDER_COMMON,
+    ...keys("environment", "defaultPageSize", "baseUrl"),
+});
 /**
  * A maintained Overflow connection.
  *
@@ -1501,7 +1452,21 @@ ${accountInstructions
  * cannot infer — Overflow's keys carry no recognizable environment marker — so
  * the operator states it, and the title and the guide's first line repeat it.
  */
-export function overflow(id, options) {
+export const overflow = asProviderFactory({
+    name: "overflow",
+    title: "Overflow",
+    kind: "api",
+    readme: "Overflow",
+    bundle: {
+        "baselineGzip": 20410,
+        "maxGzip": 80410,
+        "note": "./providers/overflow starts at 20,410 B gzip: a hand-written api() surface over the guarded transport with no hosted-MCP interface, so it pulls in neither remoteMcp nor the Effect core the other providers carry. The cap uses the existing baseline + 60,000 B policy.",
+    },
+    skill,
+    options: OVERFLOW_OPTIONS,
+    create: overflowConnector,
+});
+function overflowConnector(id, options) {
     const purpose = options.purpose?.trim();
     if (!purpose) {
         throw new Error("overflow() requires a non-empty nonprofit purpose.");
@@ -1511,9 +1476,7 @@ export function overflow(id, options) {
         throw new Error('overflow() requires environment: "production" or "staging". Overflow issues separate credentials for each, and connecta will not guess which one a deployment means.');
     }
     const defaultPageSize = options.defaultPageSize ?? DEFAULT_PAGE_SIZE;
-    if (!Number.isInteger(defaultPageSize) ||
-        defaultPageSize < 1 ||
-        defaultPageSize > MAX_PAGE_SIZE) {
+    if (!Number.isInteger(defaultPageSize) || defaultPageSize < 1 || defaultPageSize > MAX_PAGE_SIZE) {
         throw new Error(`overflow() defaultPageSize must be a whole number between 1 and ${MAX_PAGE_SIZE}.`);
     }
     const send = overflowTransport(options.baseUrl?.trim() || OVERFLOW_API_BASE_URLS[environment]);
@@ -1533,7 +1496,7 @@ export function overflow(id, options) {
                 const payload = asRecord(await callOverflow(send, { method: "GET", path: "/locations", query: { limit: 1, sortBy: "displayOrder" } }, {
                     ...ctx,
                     credential: {
-                        get: async (field) => (field ? values[field] ?? null : null),
+                        get: async (field) => (field ? (values[field] ?? null) : null),
                         getAll: async () => values,
                     },
                 }));

package/dist/providers/overflow/skill.generated.js

@@ -0,0 +1,8 @@
+// Generated by scripts/generate-providers.mjs; do not edit.
+export const skill = {
+    "content": "\n\n## Money, units, and status\n\nOverflow does not use one money unit, so read the unit per field:\n\n- Fields named `…InCents` are cents: deposits and their line items and\n  totals, refunds, chargebacks, and the authorize body.\n- Recurring-gift `amount` is dollars: Overflow's create and update bodies\n  say so, and its subscription reads carry the same field.\n- Contribution `amount` states no unit in Overflow's schema. Check one known\n  gift before reporting totals.\n- `status` is per asset type. A gift is final at `PAID_OUT` (cash),\n  `CONFIRMED` (manual cash, crypto), `CONTRIBUTION_RECEIVED` or `BILLED`\n  (stock), and `CONTRIBUTION_RECEIVED` (DAF). `statusBucket` filters by\n  Overflow's coarser PENDING / CONFIRMED / FAILED / CANCELED.\n- There is no funds endpoint. A gift reports its campaign and subcampaign;\n  `fundId` and `subFundId` appear only in the payment-authorize body.\n- Poll for changes with `minimumUpdatedDate`; `minimumInitiatedDate` is\n  when the donor gave.\n\n## Reconcile a deposit\n\n`list_deposits` → `get_deposit_summary` for totals → `get_deposit` for the\nline items, whose `referenceId` names the contribution, refund, or\nchargeback. Contributions also carry their `depositId`.\n\n## Donor data is sensitive\n\nGiving history and donor contact details are personal data. Reduce inside\n`execute_code` — aggregate, count, or keep ids — and return names, emails,\nphones, or addresses only when the task needs them. Lists and contributions\nalready omit contact details; `get_donor` is the deliberate lookup.\n\n## Writes go through `overflow_api_mutate`\n\nEvery Overflow write moves money or changes a donor record, so none has a\nnamed tool. Each is one approval-gated `overflow_api_mutate` call with\nOverflow's documented body:\n\n- `POST /contributions` charges a saved payment method. Its `amount`\n  states no unit; confirm it against a known gift of the same donor before\n  charging. `POST /payments/authorize` takes `amountInCents` and\n  authorizes a new method (`0` only saves it).\n- `POST /contributions/{id}/initiate-refund` (under $1,000 only).\n- `POST` / `PATCH` / `DELETE /subscriptions/{donorId}[/{subscriptionId}]`\n  creates, changes, or cancels a recurring gift. Its `amount` is dollars,\n  not cents: a $50.00 monthly gift is `amount: 50`. Cancel takes a\n  `cancellationReason` body.\n- `POST /donors`, `PATCH /donors/{id}`.\n\nResolve every id first — donors, payment methods (`list_payment_methods`),\ncampaigns, locations — and never guess one. A write that fails with HTTP 5xx\nmay still have happened: read before retrying.\n\n## Hatches, pages, and limits\n\n- `overflow_api_get` reaches the reads without a named tool: single refunds,\n  chargebacks, campaigns, locations, webhooks, and recurring gifts\n  (`/subscriptions/{donorId}/{id}`), and Tap devices, groups, and\n  destinations (`/tap/devices`, `/tap/groups`, `/tap/destinations`). Paths\n  are below `/api/v3`; array filters repeat a `name[]` query pair.\n- Lists page by number: pass `page` while `page.hasMore` is true.\n  `list_webhook_event_logs` pages by cursor instead.\n- Overflow allows 120 requests per minute per API client. This connection\n  admits the same per runtime, four at a time, which approximates rather than\n  enforces it; a `rate_limited` failure carries the wait.\n",
+    "instructionsHeading": "Nonprofit instructions",
+    "fragments": {
+        "guide_0": "\n\n## Money, units, and status\n\nOverflow does not use one money unit, so read the unit per field:\n\n- Fields named `…InCents` are cents: deposits and their line items and\n  totals, refunds, chargebacks, and the authorize body.\n- Recurring-gift `amount` is dollars: Overflow's create and update bodies\n  say so, and its subscription reads carry the same field.\n- Contribution `amount` states no unit in Overflow's schema. Check one known\n  gift before reporting totals.\n- `status` is per asset type. A gift is final at `PAID_OUT` (cash),\n  `CONFIRMED` (manual cash, crypto), `CONTRIBUTION_RECEIVED` or `BILLED`\n  (stock), and `CONTRIBUTION_RECEIVED` (DAF). `statusBucket` filters by\n  Overflow's coarser PENDING / CONFIRMED / FAILED / CANCELED.\n- There is no funds endpoint. A gift reports its campaign and subcampaign;\n  `fundId` and `subFundId` appear only in the payment-authorize body.\n- Poll for changes with `minimumUpdatedDate`; `minimumInitiatedDate` is\n  when the donor gave.\n\n## Reconcile a deposit\n\n`list_deposits` → `get_deposit_summary` for totals → `get_deposit` for the\nline items, whose `referenceId` names the contribution, refund, or\nchargeback. Contributions also carry their `depositId`.\n\n## Donor data is sensitive\n\nGiving history and donor contact details are personal data. Reduce inside\n`execute_code` — aggregate, count, or keep ids — and return names, emails,\nphones, or addresses only when the task needs them. Lists and contributions\nalready omit contact details; `get_donor` is the deliberate lookup.\n\n## Writes go through `overflow_api_mutate`\n\nEvery Overflow write moves money or changes a donor record, so none has a\nnamed tool. Each is one approval-gated `overflow_api_mutate` call with\nOverflow's documented body:\n\n- `POST /contributions` charges a saved payment method. Its `amount`\n  states no unit; confirm it against a known gift of the same donor before\n  charging. `POST /payments/authorize` takes `amountInCents` and\n  authorizes a new method (`0` only saves it).\n- `POST /contributions/{id}/initiate-refund` (under $1,000 only).\n- `POST` / `PATCH` / `DELETE /subscriptions/{donorId}[/{subscriptionId}]`\n  creates, changes, or cancels a recurring gift. Its `amount` is dollars,\n  not cents: a $50.00 monthly gift is `amount: 50`. Cancel takes a\n  `cancellationReason` body.\n- `POST /donors`, `PATCH /donors/{id}`.\n\nResolve every id first — donors, payment methods (`list_payment_methods`),\ncampaigns, locations — and never guess one. A write that fails with HTTP 5xx\nmay still have happened: read before retrying.\n\n## Hatches, pages, and limits\n\n- `overflow_api_get` reaches the reads without a named tool: single refunds,\n  chargebacks, campaigns, locations, webhooks, and recurring gifts\n  (`/subscriptions/{donorId}/{id}`), and Tap devices, groups, and\n  destinations (`/tap/devices`, `/tap/groups`, `/tap/destinations`). Paths\n  are below `/api/v3`; array filters repeat a `name[]` query pair.\n- Lists page by number: pass `page` while `page.hasMore` is true.\n  `list_webhook_event_logs` pages by cursor instead.\n- Overflow allows 120 requests per minute per API client. This connection\n  admits the same per runtime, four at a time, which approximates rather than\n  enforces it; a `rate_limited` failure carries the wait.\n"
+    }
+};

package/dist/providers/{planning-center.d.ts → planning-center/index.d.ts}

@@ -1,4 +1,4 @@
-import type { Connector, ConnectorCallAdmissionPolicy } from "../types.js";
+import type { ConnectorCallAdmissionPolicy } from "../../types.js";
 /** Planning Center's REST origin. Override only for a proxy or test double. */
 export declare const PLANNING_CENTER_API_BASE_URL = "https://api.planningcenteronline.com";
 /** The products behind `/<app>/v2`, in the order their guides list them. */
@@ -56,4 +56,4 @@ export interface PlanningCenterOptions {
     baseUrl?: string;
 }
 /** A maintained Planning Center Online connection over the REST API. */
-export declare function planningCenter(id: string, options: PlanningCenterOptions): Connector;
+export declare const planningCenter: import("../../provider.js").ProviderFactory<PlanningCenterOptions>;