@knowttl/checkpoint-axi 0.1.2 → 0.1.3

package/README.md

@@ -59,6 +59,9 @@ Use the hook, the skill or both; neither is required for ordinary CLI use.
 Start investigations with the compact [agent guide](docs/agent-guide.md), linked from [SKILL.md](skills/checkpoint-axi/SKILL.md).
 Use the [command reference](docs/commands.md) for the full catalogue.
 
+Email search and event list support opt-in read-only detail expansion with `--expand`.
+See [bounded detail expansion](docs/reads.md#bounded-detail-expansion) for limits, shared budgets, partial results and examples.
+
 ## What it covers
 
 <!-- command-registry:start -->
@@ -73,7 +76,15 @@ Use the [command reference](docs/commands.md) for the full catalogue.
 | exception | Inspect eight kinds; preview or request supported rule changes | Read / gated write | [Guide](docs/writes.md) |
 | home | Show local profile context without resolving secrets | Local read | [Guide](docs/configuration.md) |
 | profile | List named file profiles | Local read | [Guide](docs/configuration.md) |
-| task | Inspect asynchronous action status without polling | Read | [Guide](docs/reads.md#event-and-task-reads) |
+| task | Inspect asynchronous action status once or with bounded read-only waiting | Read | [Guide](docs/reads.md#event-and-task-reads) |
+
+Task show reads once by default; use task show <id> --wait --timeout 30 for bounded read-only polling.
+--timeout is required with --wait (integer seconds, range 1-60); --poll-interval defaults to 2 seconds (range 1-10), and --wait-requests defaults to 10 total HTTP requests (range 2-10), including auth and retries.
+Wait stops on completed, failed or stopped; init, inprogress and paused remain non-terminal.
+The task snapshot contains the final status; wait reports terminal, polls (validated status snapshots) and reason (terminal, timeout or budget).
+A non-terminal stop exits 0 with the last snapshot and a read-only next-step hint; inspect wait even on exit 0 and include task,wait,help with --fields.
+Timeout covers auth, requests, retries and delays within the shared 8 MiB response cap; failures before any snapshot remain errors.
+Waiting never resubmits an action or changes remote state; a failed or stopped task is terminal, not successful execution.
 
 Email search --filter accepts a JSON array of {saasAttrName, saasAttrOp, saasAttrValue}.
 saasAttrName is a field path, not a closed enum (guide pages 19-20).

package/dist/src/catalogue.d.ts

@@ -1,5 +1,7 @@
 export declare const DESCRIPTION = "Inspect Check Point Email Security through compact, guarded AXI commands";
 export declare const SETUP_GUIDANCE = "Run checkpoint-axi config path; configure a region, clientId and clientSecret env/file SecretRef";
+export declare const TASK_WAIT_GUIDANCE: string;
+export declare const DETAIL_EXPANSION_GUIDANCE: string;
 export declare const EMAIL_PAYLOAD_ATTRIBUTES: string[];
 export declare const EMAIL_FILTER_GUIDANCE: string;
 export declare const catalogue: {
@@ -194,6 +196,9 @@ export declare const catalogue: {
             readonly "event-ids": "Comma-separated event IDs; default unfiltered";
             readonly "confidence-indicator": "Confidence indicator selector; default unfiltered";
             readonly description: "Description selector; default unfiltered";
+            readonly expand: string;
+            readonly "expand-limit": string;
+            readonly "expand-requests": string;
             readonly "start-date": string;
             readonly "end-date": string;
             readonly limit: string;
@@ -205,7 +210,7 @@ export declare const catalogue: {
             readonly full: string;
             readonly help: string;
         };
-        readonly examples: readonly ["checkpoint-axi event list", "checkpoint-axi event list --all --event-types phishing", "checkpoint-axi event list --cursor <cursor>"];
+        readonly examples: readonly ["checkpoint-axi event list", "checkpoint-axi event list --limit 5 --expand --event-types phishing", "checkpoint-axi event list --cursor <cursor>"];
     };
     readonly "event show": {
         readonly description: "Show one event, actions taken and available actions (API guide pages 34-37)";
@@ -219,20 +224,27 @@ export declare const catalogue: {
         readonly examples: readonly ["checkpoint-axi event show <id>", "checkpoint-axi event show <id> --full"];
     };
     readonly "task show": {
-        readonly description: "Show asynchronous task status and per-action details (API guide pages 30-33)";
+        readonly description: "Show asynchronous task status and per-action details, optionally waiting within explicit bounds (API guide pages 30-33)";
         readonly flags: {
-            config: string;
-            profile: string;
-            fields: string;
-            full: string;
-            help: string;
+            readonly wait: "Poll existing task status only; default false; requires --timeout; terminal: completed, failed, stopped";
+            readonly timeout: "Total wait timeout in integer seconds, range 1-60; required with --wait; includes auth and retries";
+            readonly "poll-interval": "Seconds between status reads; default 2, range 1-10; requires --wait";
+            readonly "wait-requests": "Total HTTP request cap including auth and retries; default 10, range 2-10; requires --wait";
+            readonly config: string;
+            readonly profile: string;
+            readonly fields: string;
+            readonly full: string;
+            readonly help: string;
         };
-        readonly examples: readonly ["checkpoint-axi task show <id>", "checkpoint-axi task show <id> --full"];
+        readonly examples: readonly ["checkpoint-axi task show <id>", "checkpoint-axi task show <id> --wait --timeout 30 --fields task,wait,help", "checkpoint-axi task show <id> --full"];
     };
     readonly "email search": {
         readonly description: "Search emails with a bounded date window and scroll continuation (API guide pages 18-27)";
         readonly flags: {
             readonly filter: `JSON array of {saasAttrName, saasAttrOp, saasAttrValue} filters. ${string}`;
+            readonly expand: string;
+            readonly "expand-limit": string;
+            readonly "expand-requests": string;
             readonly "start-date": string;
             readonly "end-date": string;
             readonly limit: string;
@@ -246,7 +258,7 @@ export declare const catalogue: {
             readonly full: string;
             readonly help: string;
         };
-        readonly examples: readonly ["checkpoint-axi email search", "checkpoint-axi email search --all --saas office365_emails", "checkpoint-axi email search --cursor <cursor>"];
+        readonly examples: readonly ["checkpoint-axi email search", "checkpoint-axi email search --limit 5 --expand --saas office365_emails", "checkpoint-axi email search --cursor <cursor>"];
     };
     readonly "email show": {
         readonly description: "Show one email with verdicts, actions and available actions (API guide pages 10-18)";

package/dist/src/catalogue.js

@@ -1,6 +1,21 @@
 import { exceptionFamilies } from "./exceptions.js";
 export const DESCRIPTION = "Inspect Check Point Email Security through compact, guarded AXI commands";
 export const SETUP_GUIDANCE = "Run checkpoint-axi config path; configure a region, clientId and clientSecret env/file SecretRef";
+export const TASK_WAIT_GUIDANCE = "Task show reads once by default; use task show <id> --wait --timeout 30 for bounded read-only polling.\n"
+    + "--timeout is required with --wait (integer seconds, range 1-60); --poll-interval defaults to 2 seconds (range 1-10), and --wait-requests defaults to 10 total HTTP requests (range 2-10), including auth and retries.\n"
+    + "Wait stops on completed, failed or stopped; init, inprogress and paused remain non-terminal.\n"
+    + "The task snapshot contains the final status; wait reports terminal, polls (validated status snapshots) and reason (terminal, timeout or budget).\n"
+    + "A non-terminal stop exits 0 with the last snapshot and a read-only next-step hint; inspect wait even on exit 0 and include task,wait,help with --fields.\n"
+    + "Timeout covers auth, requests, retries and delays within the shared 8 MiB response cap; failures before any snapshot remain errors.\n"
+    + "Waiting never resubmits an action or changes remote state; a failed or stopped task is terminal, not successful execution.";
+export const DETAIL_EXPANSION_GUIDANCE = "Email search and event list accept --expand for read-only show details in each successful row's detail field.\n"
+    + "Expansion is off by default; --expand-limit caps detail attempts (default 5, range 1-8), and --expand-requests caps total HTTP requests (default 10, range 2-10), including auth, list pages and retries.\n"
+    + "Both limit flags require --expand; listing runs first and shares the 60-second and 8 MiB transport budgets with detail reads.\n"
+    + "Inspect the expansion object's expanded, complete and notExpanded fields for every unexpanded ID and its reason: item-cap-reached, request-cap-reached, budget-exhausted or per-item-error (with an error code).\n"
+    + "Partial expansion keeps every list row and successful detail; list complete/cursor describe pagination separately.\n"
+    + "Use --fields emails,expansion or --fields events,expansion to retain expansion status; --full restores full strings after redaction.\n"
+    + "With --all, list pages can consume the entire request budget before any detail is expanded; reduce the list limit or use individual show reads for remaining IDs.\n"
+    + "Expansion never downloads files, invokes available actions or changes remote state.";
 export const EMAIL_PAYLOAD_ATTRIBUTES = ["subject", "fromEmail", "attachmentCount", "internetMessageId", "received", "size", "emailLinks", "attachments", "mode", "recipients", "fromDomain", "fromUser", "fromName", "to", "toUser", "cc", "ccUser", "bcc", "bccUser", "replyToEmail", "replyToNickname", "isRead", "isDeleted", "isIncoming", "isInternal", "isOutgoing", "isQuarantined", "isQuarantineNotification", "isRestored", "isRestoreRequested", "isRestoreDeclined", "saasSpamVerdict", "SpfResult"].map((name) => `entityPayload.${name}`);
 export const EMAIL_FILTER_GUIDANCE = "saasAttrName is a field path, not a closed enum (guide pages 19-20). Documented examples: "
     + EMAIL_PAYLOAD_ATTRIBUTES.slice(0, 3).join(", ") + ". Other documented email payload paths (pages 15-16, 24-25): "
@@ -21,6 +36,11 @@ const paging = {
     cursor: "Opaque continuation bound to profile, region, operation and selectors; repeat original selectors",
 };
 const exceptionType = `Required for ${["malware", "url", "dlp"].map((kind) => `${kind}: ${exceptionFamilies[kind].types.join(",")}`).join("; ")}`;
+const expansion = {
+    expand: "Opt in to read-only show details on returned rows; default false; partial expansion is reported separately",
+    "expand-limit": "Maximum detail attempts; default 5, range 1-8; requires --expand",
+    "expand-requests": "Total HTTP request cap including auth, list pages, detail and retries; default 10, range 2-10; requires --expand",
+};
 const writeFlags = { ...common,
     ids: "Required comma-separated unique entity IDs; maximum 20",
     execute: "Send once after all write gates; default false (preview)",
@@ -70,7 +90,7 @@ export const catalogue = {
     "exception show": { description: "Show an exception by kind and ID/value; anomaly refuses ENDPOINT_UNVERIFIED pending commissioning (guide pages 52-123)",
         flags: { ...common, type: exceptionType, items: "Click-time item route instead of list route; default false" },
         examples: ["checkpoint-axi exception show phishing-block <id>", "checkpoint-axi exception show url <value> --type allow-url", "checkpoint-axi exception show click-time <id> --items --full"] },
-    "event list": { description: "List security events with bounded scroll continuation (API guide pages 37-42)", flags: { ...common, ...paging,
+    "event list": { description: "List security events with bounded scroll continuation (API guide pages 37-42)", flags: { ...common, ...paging, ...expansion,
             saas: "Comma-separated SaaS overrides; default selected profile saas or office365_emails",
             "event-types": "Comma-separated event types; default unfiltered",
             "event-states": "Comma-separated event states; default unfiltered",
@@ -78,17 +98,21 @@ export const catalogue = {
             "event-ids": "Comma-separated event IDs; default unfiltered",
             "confidence-indicator": "Confidence indicator selector; default unfiltered",
             description: "Description selector; default unfiltered",
-        }, examples: ["checkpoint-axi event list", "checkpoint-axi event list --all --event-types phishing", "checkpoint-axi event list --cursor <cursor>"] },
+        }, examples: ["checkpoint-axi event list", "checkpoint-axi event list --limit 5 --expand --event-types phishing", "checkpoint-axi event list --cursor <cursor>"] },
     "event show": { description: "Show one event, actions taken and available actions (API guide pages 34-37)", flags: common,
         examples: ["checkpoint-axi event show <id>", "checkpoint-axi event show <id> --full"] },
-    "task show": { description: "Show asynchronous task status and per-action details (API guide pages 30-33)", flags: common,
-        examples: ["checkpoint-axi task show <id>", "checkpoint-axi task show <id> --full"] },
+    "task show": { description: "Show asynchronous task status and per-action details, optionally waiting within explicit bounds (API guide pages 30-33)", flags: { ...common,
+            wait: "Poll existing task status only; default false; requires --timeout; terminal: completed, failed, stopped",
+            timeout: "Total wait timeout in integer seconds, range 1-60; required with --wait; includes auth and retries",
+            "poll-interval": "Seconds between status reads; default 2, range 1-10; requires --wait",
+            "wait-requests": "Total HTTP request cap including auth and retries; default 10, range 2-10; requires --wait",
+        }, examples: ["checkpoint-axi task show <id>", "checkpoint-axi task show <id> --wait --timeout 30 --fields task,wait,help", "checkpoint-axi task show <id> --full"] },
     "email search": { description: "Search emails with a bounded date window and scroll continuation (API guide pages 18-27)", flags: { ...common,
             saas: "SaaS override; default selected profile saas or office365_emails",
             "saas-entity": "Optional SaaS entity type selector",
-            ...paging,
+            ...paging, ...expansion,
             filter: `JSON array of {saasAttrName, saasAttrOp, saasAttrValue} filters. ${EMAIL_FILTER_GUIDANCE}`,
-        }, examples: ["checkpoint-axi email search", "checkpoint-axi email search --all --saas office365_emails", "checkpoint-axi email search --cursor <cursor>"] },
+        }, examples: ["checkpoint-axi email search", "checkpoint-axi email search --limit 5 --expand --saas office365_emails", "checkpoint-axi email search --cursor <cursor>"] },
     "email show": { description: "Show one email with verdicts, actions and available actions (API guide pages 10-18)", flags: common,
         examples: ["checkpoint-axi email show <id>", "checkpoint-axi email show <id> --full"] },
     doctor: { description: "Check regional API authentication and one bounded email search without printing tenant data", flags: { ...common, saas: "SaaS override; default selected profile saas or office365_emails" },

package/dist/src/local-commands.js

@@ -38,7 +38,7 @@ export async function localCommand(argv, store, transport) {
             const label = flag.split("=")[0];
             usage(/^--[A-Za-z][A-Za-z0-9-]*$/.test(label) ? `Unknown or duplicate flag ${label}` : undefined);
         }
-        if (["--help", "--full", "--all", "--items", "--specific", "--original", "--execute", "--dry-run"].includes(flag))
+        if (["--help", "--full", "--all", "--expand", "--wait", "--items", "--specific", "--original", "--execute", "--dry-run"].includes(flag))
             flags.set(flag, "true");
         else {
             const value = argv[++index];
@@ -50,11 +50,13 @@ export async function localCommand(argv, store, transport) {
     if (flags.has("--help"))
         return { command: name, ...entry };
     const available = isException ? name === "exception list" ? ["exceptions", "count", "total", "complete", "reason", "responseSchema", "help"] : ["exception", "responseSchema"]
-        : name === "email search" || name === "event list" ? [name === "event list" ? "events" : "emails", "count", "total", "complete", "reason", "cursor", "help"] : name.endsWith(" show") ? [name.split(" ")[0]]
+        : name === "email search" || name === "event list" ? [name === "event list" ? "events" : "emails", "count", "total", "complete", "reason", "cursor", "expansion", "help"] : name.endsWith(" show") ? [name.split(" ")[0]]
             : name === "doctor" ? ["region", "saas", "checks", "pass", "help"] : name === "auth status" ? ["profile", "source", "region", "origin", "saas", "clientSecret", "readOnly"]
                 : name === "profile list" ? ["profiles", "count", "config"] : name === "config path" ? ["config", "state"]
                     : ["bin", "description", "status", "profiles", "count", "help"];
     const isWrite = name === "email quarantine" || name === "email restore" || name === "event action" || isExceptionWrite;
+    if (name === "task show")
+        available.push("wait", "help");
     if (isWrite)
         available.splice(0, available.length, "outcome", "preview", "auditId", "tasks", "response", "responseSchema", "help");
     if (isDownload)

package/dist/src/reads.js

@@ -1,5 +1,6 @@
 import { AxiError } from "axi-sdk-js";
 import { createHash } from "node:crypto";
+import { setTimeout as delay } from "node:timers/promises";
 import { z } from "zod";
 import { REGIONS } from "./regions.js";
 import { gateway, READ_LIMITS } from "./transport.js";
@@ -30,12 +31,32 @@ export async function readCommand(name, id, flags, store, transport) {
     const isEvent = name.startsWith("event ");
     if (name.endsWith(" show") && (!id || !/^[A-Za-z0-9_-]+$/.test(id)))
         invalid(`${name} requires one ID containing letters, digits, underscores or hyphens`);
+    const waiting = flags.has("--wait");
+    if (!waiting && ["--timeout", "--poll-interval", "--wait-requests"].some((flag) => flags.has(flag)))
+        invalid("Wait limits require --wait");
+    const timeoutText = flags.get("--timeout") ?? "";
+    const intervalText = flags.get("--poll-interval") ?? "2";
+    const waitRequestsText = flags.get("--wait-requests") ?? "10";
+    if (waiting && (!/^[1-9][0-9]*$/.test(timeoutText) || Number(timeoutText) > 60))
+        invalid("--wait requires --timeout in integer seconds from 1 to 60");
+    if (!/^(?:[1-9]|10)$/.test(intervalText))
+        invalid("--poll-interval must be integer seconds from 1 to 10");
+    if (!/^(?:[2-9]|10)$/.test(waitRequestsText))
+        invalid("--wait-requests must be an integer from 2 to 10");
     const limitText = flags.get("--limit") ?? "50";
     if (!/^[1-9][0-9]*$/.test(limitText) || Number(limitText) > READ_LIMITS.rows)
         invalid("--limit must be an integer from 1 to 1000");
     if (flags.has("--all") && flags.has("--limit"))
         invalid("Use either --all or --limit");
     const limit = flags.has("--all") ? READ_LIMITS.rows : Number(limitText);
+    if (!flags.has("--expand") && (flags.has("--expand-limit") || flags.has("--expand-requests")))
+        invalid("Expansion limits require --expand");
+    const expandLimitText = flags.get("--expand-limit") ?? "5";
+    const expandRequestsText = flags.get("--expand-requests") ?? "10";
+    if (!/^[1-8]$/.test(expandLimitText))
+        invalid("--expand-limit must be an integer from 1 to 8");
+    if (!/^(?:[2-9]|10)$/.test(expandRequestsText))
+        invalid("--expand-requests must be an integer from 2 to 10");
     let extended;
     if (flags.has("--filter")) {
         let raw;
@@ -100,8 +121,14 @@ export async function readCommand(name, id, flags, store, transport) {
         origin: REGIONS[profile.region], operation: name, ...(isEvent ? { selectors } : { entityFilter, extended }) })).digest("hex");
     if (cursor && cursor.binding !== binding)
         invalid("Cursor does not match this profile, region or search selectors; repeat the original selectors");
-    const api = gateway(profile, store.credentials(profile, true), store.redactor, transport);
-    if (name.endsWith(" show")) {
+    const now = transport?.now ?? Date.now;
+    const wait = transport?.wait ?? delay;
+    const timeout = waiting ? Number(timeoutText) * 1_000 : READ_LIMITS.milliseconds;
+    const deadline = now() + timeout;
+    const api = gateway(profile, store.credentials(profile, true), store.redactor, transport, false, Number(waiting ? waitRequestsText : expandRequestsText), timeout);
+    async function detail(id) {
+        if (!/^[A-Za-z0-9_-]+$/.test(id))
+            throw new AxiError("Invalid detail ID", "RESPONSE_INVALID", []);
         const kind = name.split(" ")[0];
         const page = await api.read({ path: name === "task show" ? `/task/${id}` : isEvent ? `/event/${id}` : `/search/entity/${id}` });
         const response = { responseData: page.dataIsArray ? page.records : page.records[0] };
@@ -117,8 +144,43 @@ export async function readCommand(name, id, flags, store, transport) {
         const returnedId = name === "task show" ? String(record.id) : isEvent ? record.eventId : record.entityInfo?.entityId;
         if (returnedId !== id)
             throw responseInvalid(`Invalid ${kind} detail response`, response, [[...recordPath, ...(name === "task show" ? ["id"] : isEvent ? ["eventId"] : ["entityInfo", "entityId"])]]);
-        return { [kind]: parsed.data };
+        return parsed.data;
     }
+    if (name === "task show" && waiting) {
+        let snapshot;
+        let polls = 0;
+        let reason;
+        for (;;) {
+            try {
+                snapshot = await detail(id);
+                polls++;
+            }
+            catch (error) {
+                if (!snapshot || !(error instanceof AxiError) || !["REQUEST_BUDGET", "TIME_BUDGET", "RESPONSE_TOO_LARGE"].includes(error.code))
+                    throw error;
+                reason = error.code === "TIME_BUDGET" ? "timeout" : "budget";
+                break;
+            }
+            if (["completed", "failed", "stopped"].includes(snapshot.status)) {
+                reason = "terminal";
+                break;
+            }
+            const remaining = deadline - now();
+            if (remaining <= 0) {
+                reason = "timeout";
+                break;
+            }
+            await wait(Math.min(Number(intervalText) * 1_000, remaining));
+            if (now() >= deadline) {
+                reason = "timeout";
+                break;
+            }
+        }
+        return { task: snapshot, wait: { terminal: reason === "terminal", polls, reason },
+            ...(reason === "terminal" ? {} : { help: [`Run checkpoint-axi task show ${id}${flags.has("--config") ? " --config <absolute-path>" : ""}${flags.has("--profile") ? " --profile <name>" : ""} to read status again; the task may still run, never resubmit its action to obtain status`] }) };
+    }
+    if (name.endsWith(" show"))
+        return { [name.split(" ")[0]]: await detail(id) };
     const results = [];
     let scroll = cursor?.scroll;
     let offset = cursor?.offset ?? 0;
@@ -171,7 +233,43 @@ export async function readCommand(name, id, flags, store, transport) {
     }
     const next = exhausted ? undefined : Buffer.from(JSON.stringify({ version: 1, binding, start, end, scroll, offset,
         ...(replayRisk ? { replayRisk: true } : {}) })).toString("base64url");
+    let expansion;
+    if (flags.has("--expand")) {
+        let attempted = 0;
+        let expanded = 0;
+        let stopped;
+        const notExpanded = [];
+        for (const row of results) {
+            const itemId = row.id;
+            if (stopped) {
+                notExpanded.push({ id: itemId, ...stopped });
+                continue;
+            }
+            if (attempted >= Number(expandLimitText)) {
+                notExpanded.push({ id: itemId, reason: "item-cap-reached" });
+                continue;
+            }
+            attempted++;
+            try {
+                row.detail = await detail(itemId);
+                expanded++;
+            }
+            catch (error) {
+                if (!(error instanceof AxiError))
+                    throw error;
+                const code = error.code;
+                if (["REQUEST_BUDGET", "TIME_BUDGET", "RESPONSE_TOO_LARGE"].includes(code)) {
+                    stopped = { reason: code === "REQUEST_BUDGET" ? "request-cap-reached" : "budget-exhausted", code };
+                    notExpanded.push({ id: itemId, ...stopped });
+                }
+                else
+                    notExpanded.push({ id: itemId, reason: "per-item-error", code });
+            }
+        }
+        expansion = { expanded, attempted, itemLimit: Number(expandLimitText), requestLimit: Number(expandRequestsText), complete: notExpanded.length === 0, notExpanded };
+    }
     return { [isEvent ? "events" : "emails"]: results, count: results.length, ...(total === undefined ? {} : { total }), complete: exhausted && !replayRisk,
+        ...(expansion ? { expansion } : {}),
         ...(replayRisk ? { reason: "replay-may-skip-or-repeat" } : {}), ...(next ? { cursor: next } : {}),
         help: [...(results.length ? [isEvent ? "Run checkpoint-axi event show <id> for event details and available actions" : "Run checkpoint-axi email show <id> for verdicts, actions and available actions"] : [`No ${isEvent ? "events" : "emails"} matched this search window and selectors`]),
             ...(next ? ["Repeat this search with --cursor <cursor> and the original selectors to continue"] : []),

package/dist/src/transport.d.ts

@@ -72,7 +72,7 @@ export declare function decodePage(value: unknown, unverified?: boolean): Page;
 export declare function gateway(profile: SelectedProfile, credentials: {
     clientId: string;
     clientSecret: string;
-}, redactor: SecretRedactor, dependencies?: Dependencies, mutation?: boolean): {
+}, redactor: SecretRedactor, dependencies?: Dependencies, mutation?: boolean, requestLimit?: number, milliseconds?: number): {
     authenticate: () => Promise<void>;
     read: (query: ReadRequest) => Promise<Page>;
     download: (query: DownloadRequest) => Promise<{

package/dist/src/transport.js

@@ -66,10 +66,10 @@ function tokenFrom(value, redactor) {
         throw failure("AUTH_SHAPE_UNVERIFIED", "Unverified auth token response shape; owner commissioning is required (API guide page 8 does not specify the shape)");
     return [...tokens][0];
 }
-export function gateway(profile, credentials, redactor, dependencies = {}, mutation = false) {
+export function gateway(profile, credentials, redactor, dependencies = {}, mutation = false, requestLimit = READ_LIMITS.requests, milliseconds = READ_LIMITS.milliseconds) {
     const now = dependencies.now ?? Date.now;
     const wait = dependencies.wait ?? delay;
-    const deadline = now() + READ_LIMITS.milliseconds;
+    const deadline = now() + Math.min(milliseconds, READ_LIMITS.milliseconds);
     let bytes = 0;
     let requests = 0;
     let token;
@@ -86,7 +86,7 @@ export function gateway(profile, credentials, redactor, dependencies = {}, mutat
         const remaining = deadline - now();
         if (remaining <= 0)
             throw failure("TIME_BUDGET", "Read time budget exhausted");
-        if (++requests > (mutation ? 44 : READ_LIMITS.requests))
+        if (++requests > (mutation ? 44 : Math.min(requestLimit, READ_LIMITS.requests)))
             throw failure("REQUEST_BUDGET", "Request budget exhausted");
         const payload = body === undefined ? undefined : JSON.stringify(body);
         if (bearer)

package/docs/agent-guide.md

@@ -43,12 +43,13 @@ saasAttrOp: is, contains, startsWith, isEmpty, isNot, notContains, isNotEmpty, g
 The guide defines operators globally, without a per-field compatibility matrix; examples pair fromEmail with is and attachmentCount with greaterThan.
 saasAttrValue is a string.
 Event selectors include --event-types, --event-states, --severities, --event-ids (comma-separated), --confidence-indicator and --description; --saas accepts comma-separated SaaS values here.
-Event detail includes available actions; task detail reads status and per-action results once, without waiting or polling:
+Event detail includes available actions; task detail reads status and per-action results once by default:
 
 ```sh
 checkpoint-axi event list --all --event-types phishing --fields events,count,total,complete,reason,cursor
 checkpoint-axi event show synthetic-event-1 --fields event
 checkpoint-axi task show synthetic-task-1 --fields task
+checkpoint-axi task show synthetic-task-1 --wait --timeout 30 --fields task,wait,help
 ```
 
 Inspect exception rules with exception list <kind> or exception show <kind> <id-or-value>:
@@ -74,7 +75,27 @@ Inspect complete, reason and cursor even on exit 0; limits and page replay can l
 Resume with --cursor <returned-cursor> and the original selectors/profile; see [continuation limitations](reads.md#email-reads).
 Exception --all bounds returned rows at 1000 but cannot fetch additional pages: no continuation parameter is documented.
 Email/event detail already includes available actions, avoiding a separate action lookup.
-There is no bulk detail expansion, cross-resource pivot or task wait flag; use returned IDs for individual reads and never replay an action to obtain its status.
+Email search and event list accept --expand for read-only show details in each successful row's detail field.
+Expansion is off by default; --expand-limit caps detail attempts (default 5, range 1-8), and --expand-requests caps total HTTP requests (default 10, range 2-10), including auth, list pages and retries.
+Both limit flags require --expand; listing runs first and shares the 60-second and 8 MiB transport budgets with detail reads.
+Inspect the expansion object's expanded, complete and notExpanded fields for every unexpanded ID and its reason: item-cap-reached, request-cap-reached, budget-exhausted or per-item-error (with an error code).
+Partial expansion keeps every list row and successful detail; list complete/cursor describe pagination separately.
+Use --fields emails,expansion or --fields events,expansion to retain expansion status; --full restores full strings after redaction.
+With --all, list pages can consume the entire request budget before any detail is expanded; reduce the list limit or use individual show reads for remaining IDs.
+Expansion never downloads files, invokes available actions or changes remote state.
+Task show reads once by default; use task show <id> --wait --timeout 30 for bounded read-only polling.
+--timeout is required with --wait (integer seconds, range 1-60); --poll-interval defaults to 2 seconds (range 1-10), and --wait-requests defaults to 10 total HTTP requests (range 2-10), including auth and retries.
+Wait stops on completed, failed or stopped; init, inprogress and paused remain non-terminal.
+The task snapshot contains the final status; wait reports terminal, polls (validated status snapshots) and reason (terminal, timeout or budget).
+A non-terminal stop exits 0 with the last snapshot and a read-only next-step hint; inspect wait even on exit 0 and include task,wait,help with --fields.
+Timeout covers auth, requests, retries and delays within the shared 8 MiB response cap; failures before any snapshot remain errors.
+Waiting never resubmits an action or changes remote state; a failed or stopped task is terminal, not successful execution.
+There is no cross-resource pivot; use returned IDs for individual reads and never replay an action to obtain its status.
+
+```sh
+checkpoint-axi email search --limit 5 --expand --expand-limit 3 --expand-requests 6 --fields emails,count,complete,cursor,expansion
+checkpoint-axi event list --limit 5 --expand --fields events,count,complete,cursor,expansion
+```
 
 --fields selects top-level output keys, not nested row columns: request emails or events together with count,total,complete,reason,cursor for lists, and email,event,task or exception for details.
 Email rows contain id,subject,state,verdict; event rows contain id,type,state,severity; documented exception rows contain id,sender,subject,comment.

package/docs/commands.md

@@ -32,7 +32,7 @@ Run each command with `--help` for concise help.
 | `checkpoint-axi exception update` | Preview or replace one exception; phishing/spam/click-time require an ID; malware/url/dlp identify it in --input without an ID; anomaly unsupported (guide pages 64-67, 78-81, 94-96, 105-107, 116-117, 126-127) |
 | `checkpoint-axi home` | Show local profile context without resolving secrets or making API calls |
 | `checkpoint-axi profile list` | List file profiles without resolving secrets; --profile does not filter this list |
-| `checkpoint-axi task show` | Show asynchronous task status and per-action details (API guide pages 30-33) |
+| `checkpoint-axi task show` | Show asynchronous task status and per-action details, optionally waiting within explicit bounds (API guide pages 30-33) |
 
 ## attachment download
 
@@ -197,11 +197,14 @@ Search emails with a bounded date window and scroll continuation (API guide page
 | `--limit` | Maximum rows; default 50, range 1-1000; cannot combine with --all |
 | `--all` | Read up to 1000 rows / 10 HTTP requests / 60 seconds / 8 MiB; default false |
 | `--cursor` | Opaque continuation bound to profile, region, operation and selectors; repeat original selectors |
+| `--expand` | Opt in to read-only show details on returned rows; default false; partial expansion is reported separately |
+| `--expand-limit` | Maximum detail attempts; default 5, range 1-8; requires --expand |
+| `--expand-requests` | Total HTTP request cap including auth, list pages, detail and retries; default 10, range 2-10; requires --expand |
 | `--filter` | JSON array of {saasAttrName, saasAttrOp, saasAttrValue} filters. saasAttrName is a field path, not a closed enum (guide pages 19-20). Documented examples: entityPayload.subject, entityPayload.fromEmail, entityPayload.attachmentCount. Other documented email payload paths (pages 15-16, 24-25): entityPayload.internetMessageId, entityPayload.received, entityPayload.size, entityPayload.emailLinks, entityPayload.attachments, entityPayload.mode, entityPayload.recipients, entityPayload.fromDomain, entityPayload.fromUser, entityPayload.fromName, entityPayload.to, entityPayload.toUser, entityPayload.cc, entityPayload.ccUser, entityPayload.bcc, entityPayload.bccUser, entityPayload.replyToEmail, entityPayload.replyToNickname, entityPayload.isRead, entityPayload.isDeleted, entityPayload.isIncoming, entityPayload.isInternal, entityPayload.isOutgoing, entityPayload.isQuarantined, entityPayload.isQuarantineNotification, entityPayload.isRestored, entityPayload.isRestoreRequested, entityPayload.isRestoreDeclined, entityPayload.saasSpamVerdict, entityPayload.SpfResult. saasAttrOp: is, contains, startsWith, isEmpty, isNot, notContains, isNotEmpty, greaterThan, lessThan. The guide defines operators globally, without a per-field compatibility matrix; examples pair fromEmail with is and attachmentCount with greaterThan. saasAttrValue is a string. |
 
 ```sh
 checkpoint-axi email search
-checkpoint-axi email search --all --saas office365_emails
+checkpoint-axi email search --limit 5 --expand --saas office365_emails
 checkpoint-axi email search --cursor <cursor>
 ```
 
@@ -260,6 +263,9 @@ List security events with bounded scroll continuation (API guide pages 37-42)
 | `--limit` | Maximum rows; default 50, range 1-1000; cannot combine with --all |
 | `--all` | Read up to 1000 rows / 10 HTTP requests / 60 seconds / 8 MiB; default false |
 | `--cursor` | Opaque continuation bound to profile, region, operation and selectors; repeat original selectors |
+| `--expand` | Opt in to read-only show details on returned rows; default false; partial expansion is reported separately |
+| `--expand-limit` | Maximum detail attempts; default 5, range 1-8; requires --expand |
+| `--expand-requests` | Total HTTP request cap including auth, list pages, detail and retries; default 10, range 2-10; requires --expand |
 | `--saas` | Comma-separated SaaS overrides; default selected profile saas or office365_emails |
 | `--event-types` | Comma-separated event types; default unfiltered |
 | `--event-states` | Comma-separated event states; default unfiltered |
@@ -270,7 +276,7 @@ List security events with bounded scroll continuation (API guide pages 37-42)
 
 ```sh
 checkpoint-axi event list
-checkpoint-axi event list --all --event-types phishing
+checkpoint-axi event list --limit 5 --expand --event-types phishing
 checkpoint-axi event list --cursor <cursor>
 ```
 
@@ -473,7 +479,7 @@ checkpoint-axi profile list --config /example/config.json
 
 ## task show
 
-Show asynchronous task status and per-action details (API guide pages 30-33)
+Show asynchronous task status and per-action details, optionally waiting within explicit bounds (API guide pages 30-33)
 
 | Flag | Meaning |
 |---|---|
@@ -482,8 +488,13 @@ Show asynchronous task status and per-action details (API guide pages 30-33)
 | `--fields` | Comma-separated output fields; default all |
 | `--full` | Show full strings instead of 500-character previews; default false |
 | `--help` | Show concise command help; default false |
+| `--wait` | Poll existing task status only; default false; requires --timeout; terminal: completed, failed, stopped |
+| `--timeout` | Total wait timeout in integer seconds, range 1-60; required with --wait; includes auth and retries |
+| `--poll-interval` | Seconds between status reads; default 2, range 1-10; requires --wait |
+| `--wait-requests` | Total HTTP request cap including auth and retries; default 10, range 2-10; requires --wait |
 
 ```sh
 checkpoint-axi task show <id>
+checkpoint-axi task show <id> --wait --timeout 30 --fields task,wait,help
 checkpoint-axi task show <id> --full
 ```

package/docs/reads.md

@@ -54,7 +54,21 @@ These diagnostics support owner commissioning; the actual failing fields behind
 `checkpoint-axi event list` returns compact ID, type, state and severity rows (guide pages 37-42).
 `checkpoint-axi event show <id>` preserves event details, action history and available actions (guide pages 34-37).
 `checkpoint-axi task show <id>` returns task status, progress and per-action details for asynchronous actions (guide pages 30-33).
-Task states are `init`, `inprogress`, `completed`, `failed`, `stopped` and `paused`; task show performs one read without polling.
+Task states are `init`, `inprogress`, `completed`, `failed`, `stopped` and `paused`; task show performs one read by default.
+Use `checkpoint-axi task show <id> --wait --timeout 30` to poll an already submitted task without resubmitting its action.
+`--wait` requires an explicit `--timeout` in integer seconds, from 1 to 60.
+`--poll-interval` defaults to 2 seconds, from 1 to 10; `--wait-requests` defaults to 10 total HTTP requests, from 2 to 10, including authentication, re-exchange and retries.
+These limits require `--wait` and invalid bounds fail before configuration or network access.
+The timeout includes authentication, requests, retries and delays, and the existing 8 MiB response cap remains shared across polls.
+The first read starts immediately; subsequent reads wait the configured interval after the preceding read, without starting another read at or after the deadline.
+`completed`, `failed` and `stopped` are terminal; `init`, `inprogress` and `paused` remain non-terminal.
+The final `task` snapshot retains status, progress and per-action details; the `wait` object reports `terminal`, `polls` (validated status snapshots, excluding authentication and retries), and `reason`: `terminal`, `timeout` or `budget`.
+Request or response-byte exhaustion reports `budget`; an exhausted deadline or a retry delay that cannot fit reports `timeout`.
+A non-terminal stop exits 0, preserves the last snapshot and includes a `help` hint for another status read.
+Inspect `wait` even on exit 0; terminal failure or stopping does not mean the submitted action succeeded.
+Failures before any snapshot and non-budget read failures remain structured errors rather than reporting stale data as current.
+Use `--fields task,wait,help` to retain wait metadata and recovery guidance when projecting output.
+Waiting only repeats the existing task status GET; it never invokes an available action, downloads a file or resubmits the original mutation.
 Event list accepts comma-separated `--event-types`, `--event-states`, `--severities`, `--event-ids` and `--saas`, plus `--confidence-indicator` and `--description` text selectors.
 SaaS defaults to the selected profile's value or `office365_emails`; other event selectors default to unfiltered.
 Event list uses the same date windows, row and transport budgets, bound continuation and honest page replay flags as [email search](reads.md#email-reads).
@@ -63,6 +77,34 @@ Event and task text is redacted before 500-character previews; use `--full` for
 For writes, see [event actions](writes.md#event-actions).
 These reads are verified offline with synthetic fixtures; authentication, token lifetime and live API behavior require manual verification.
 
+## Bounded detail expansion
+
+Email search and event list accept `--expand` to inspect returned items through the same validated read-only path as their individual show commands.
+Expansion is off by default; available actions are returned as data and are never invoked.
+
+```sh
+checkpoint-axi email search --limit 5 --expand --expand-limit 3 --expand-requests 6
+checkpoint-axi event list --limit 5 --expand --fields events,count,complete,cursor,expansion
+```
+
+`--expand-limit` caps detail attempts, including failed attempts: default 5, minimum 1, maximum 8.
+`--expand-requests` caps total HTTP requests for the invocation: default 10, minimum 2, maximum 10.
+Both limit flags require `--expand`; invalid bounds are rejected before configuration or network access.
+Authentication, token re-exchange, list pages, detail requests and read retries all consume this request cap.
+Listing runs first; detail reads reuse its token, deadline and response-byte accounting without resetting the 60-second or 8 MiB budgets.
+Using `--all` can therefore consume the entire budget before any detail is fetched.
+Use a smaller list limit to leave room for detail, or inspect remaining IDs with individual show commands.
+
+Each successfully expanded row gains `detail`, containing the same validated payload and actions as show, with normal secret redaction and 500-character string previews (`--full` restores full strings).
+Every list row and successful detail survives expansion failures; partial expansion exits 0.
+The `expansion` object reports `expanded`, `attempted`, `itemLimit`, `requestLimit`, `complete` and `notExpanded`.
+Each `notExpanded` entry includes the item `id` and `reason`: `item-cap-reached`, `request-cap-reached`, `budget-exhausted` or `per-item-error`.
+Budget and per-item failures also include a safe error `code`; raw upstream error bodies are never printed.
+Budget exhaustion stops further detail attempts; an individual HTTP, response-validation or other read failure records that item and continues within the caps.
+List `complete`, `reason` and `cursor` still describe pagination; `complete` inside `expansion` independently describes whether every returned row was expanded.
+With `--fields`, include `expansion` alongside `emails` or `events` so omissions remain visible.
+Expansion performs no writes, actions, downloads, cross-resource pivots or task waiting.
+
 ## Exception reads
 
 `checkpoint-axi exception list <kind>` and `exception show <kind> <id-or-value>` cover `phishing-allow`, `phishing-block`, `spam-allow`, `anomaly`, `click-time`, `malware`, `url` and `dlp` (guide pages 49-123).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@knowttl/checkpoint-axi",
-  "version": "0.1.2",
+  "version": "0.1.3",
   "description": "Inspect Check Point Email Security through compact, guarded AXI commands",
   "license": "MIT",
   "type": "module",

package/skills/checkpoint-axi/SKILL.md

@@ -49,6 +49,14 @@ The guide defines operators globally, without a per-field compatibility matrix;
 saasAttrValue is a string.
 See [email-read guidance](../../docs/reads.md#email-reads) for continuation, completion status, cursor safety and page replay limitations.
 Email show <id> includes payload, verdicts, actions taken and available actions (guide pages 10-18).
+Email search and event list accept --expand for read-only show details in each successful row's detail field.
+Expansion is off by default; --expand-limit caps detail attempts (default 5, range 1-8), and --expand-requests caps total HTTP requests (default 10, range 2-10), including auth, list pages and retries.
+Both limit flags require --expand; listing runs first and shares the 60-second and 8 MiB transport budgets with detail reads.
+Inspect the expansion object's expanded, complete and notExpanded fields for every unexpanded ID and its reason: item-cap-reached, request-cap-reached, budget-exhausted or per-item-error (with an error code).
+Partial expansion keeps every list row and successful detail; list complete/cursor describe pagination separately.
+Use --fields emails,expansion or --fields events,expansion to retain expansion status; --full restores full strings after redaction.
+With --all, list pages can consume the entire request budget before any detail is expanded; reduce the list limit or use individual show reads for remaining IDs.
+Expansion never downloads files, invokes available actions or changes remote state.
 Use --full for strings truncated at 500 characters and --fields for top-level output projection.
 Cursor strings are never truncated; secret-derived strings are redacted before previews.
 Downloads require file-profile downloads.allow and an absolute new --out file, created exclusively with mode 0600 or a protected Windows DACL.
@@ -64,7 +72,14 @@ Event action --ids <ids> --name <action> previews by default and requires the ex
 Every target must offer the action on both reads; nonempty actionParameter objects or repeated action names refuse ENDPOINT_UNVERIFIED until live commissioning resolves the object/string contradiction (guide pages 36-37, 43).
 Use event show <id> to inspect all available actions, including deferred parameterized actions.
 Accepted actions return task IDs; use task show <taskId> to inspect asynchronous execution without replaying ambiguous outcomes.
-Event list supports documented selectors and shares email search bounds and continuation; event show preserves actions and task show reads asynchronous status without polling (guide pages 30-42).
+Event list supports documented selectors and shares email search bounds and continuation; event show preserves actions and task show reads asynchronous status (guide pages 30-42).
+Task show reads once by default; use task show <id> --wait --timeout 30 for bounded read-only polling.
+--timeout is required with --wait (integer seconds, range 1-60); --poll-interval defaults to 2 seconds (range 1-10), and --wait-requests defaults to 10 total HTTP requests (range 2-10), including auth and retries.
+Wait stops on completed, failed or stopped; init, inprogress and paused remain non-terminal.
+The task snapshot contains the final status; wait reports terminal, polls (validated status snapshots) and reason (terminal, timeout or budget).
+A non-terminal stop exits 0 with the last snapshot and a read-only next-step hint; inspect wait even on exit 0 and include task,wait,help with --fields.
+Timeout covers auth, requests, retries and delays within the shared 8 MiB response cap; failures before any snapshot remain errors.
+Waiting never resubmits an action or changes remote state; a failed or stopped task is terminal, not successful execution.
 Exception list/show accept phishing-allow, phishing-block, spam-allow, anomaly, click-time, malware, url and dlp (guide pages 49-123).
 For malware, url and dlp, --type is required; run exception list --help for the documented values.
 Click-time --items selects item routes; by default list routes are used.
@@ -105,4 +120,4 @@ All live-only facts remain unverified, including token response shape and lifeti
 | `npx -y @knowttl/checkpoint-axi exception update` | Preview or replace one exception; phishing/spam/click-time require an ID; malware/url/dlp identify it in --input without an ID; anomaly unsupported (guide pages 64-67, 78-81, 94-96, 105-107, 116-117, 126-127) |
 | `npx -y @knowttl/checkpoint-axi home` | Show local profile context without resolving secrets or making API calls |
 | `npx -y @knowttl/checkpoint-axi profile list` | List file profiles without resolving secrets; --profile does not filter this list |
-| `npx -y @knowttl/checkpoint-axi task show` | Show asynchronous task status and per-action details (API guide pages 30-33) |
+| `npx -y @knowttl/checkpoint-axi task show` | Show asynchronous task status and per-action details, optionally waiting within explicit bounds (API guide pages 30-33) |