@hiveku-apps/sync 0.5.5 → 0.5.6

package/README.md

@@ -11,8 +11,8 @@ Local sync CLI for a Hiveku account. Pairs with the Hiveku MCP server at
    (id, name, type, domain, owner, version, updated_at).
    It also writes the account memory (see "Account memory" below) to
    `hiveku-data/account/ACCOUNT_MEMORY.md`, read-only, and brings the folder
-   layout and memory rules in `CLAUDE.md` up to date (see "CLAUDE.md stays
-   current" below).
+   layout, the memory rules and the plan rules in `CLAUDE.md` up to date (see
+   "CLAUDE.md stays current" below).
 
 The goal: open one folder per Hiveku tenant in VS Code with Claude Code,
 with the account's memory mirrored as local `.md` files Claude can read
@@ -347,7 +347,8 @@ region `hiveku-sync codex` rewrites.
 ## CLAUDE.md stays current
 
 `init` writes `CLAUDE.md` once, and the rest of the file is yours from then on,
-except its "Folder layout" and "Memory writes" sections. They sit between
+except its "Folder layout", "Memory writes" and "Your plan and its limits"
+sections. They sit between
 `<!-- hiveku:memory:begin -->` and `<!-- hiveku:memory:end -->`, and `pull` (and
 `codex`) bring them up to date, so a folder an older release made stops teaching
 the old layout and the old `memory_update` call:
@@ -363,6 +364,12 @@ The kept-current part also states that Hiveku Memory is the source of truth,
 in the Hiveku MCP server's own words, and that the local files are a mirror
 memory wins over. `AGENTS.md` says the same.
 
+Since 0.5.6 it also carries the account's plan and its limits, in the words
+Hiveku writes once for every agent (copied byte for byte and pinned by
+sha256): call `account_usage_get` before work the plan counts, and a tool
+answer with `limit_kind` is the plan's answer, not a fault, so tell the person
+and never retry or report it. `AGENTS.md` carries the one-line version.
+
 A `CLAUDE.md` that is read-only, a link to another file, or whose markers are not
 there once each and in order is left alone (`pull` says so). `pull` never creates
 a missing one.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@hiveku-apps/sync",
-  "version": "0.5.5",
+  "version": "0.5.6",
   "description": "Sync a Hiveku account's memory/rules/skills + coder projects into a local folder, and ship projects (push/deploy/pull-env) from the terminal. Pairs with the Hiveku MCP at core.hiveku.com/mcp so Claude Code can drive Hiveku tools per-folder.",
   "type": "module",
   "license": "MIT",

package/src/claude-md.mjs

@@ -2,7 +2,8 @@
  * The part of an account folder's CLAUDE.md that hiveku-sync keeps current:
  * the folder layout and the memory-write rules, between two markers. Since
  * 2026-10-03 it also states that Hiveku Memory is the source of truth (the MCP
- * server's words) and that the local files are a mirror memory wins over.
+ * server's words) and that the local files are a mirror memory wins over, and
+ * since 0.5.6 the account's plan and its limits (the builder's words, pinned).
  *
  * `init` writes CLAUDE.md once. Everything outside the markers is the user's
  * from then on, but the layout and the memory rules change with releases
@@ -55,6 +56,42 @@ export const WORK_LOG_LINE =
   "Hiveku records this session's Doing at its first change and its Done when the session goes quiet " +
   'or ends. When the work ends, send a Done with `memory_log_add` and a one-line summary of what you ' +
   'did, if you want the log to say more than the count of changes; leave `thread` out.';
+
+/**
+ * What every agent is told about the account's plan and its limits, written once
+ * in the builder (hiveku_builder src/lib/platform-guide/plan-usage.ts, builder
+ * #1332, served with no key at /api/public/platform-guide/plan-usage) and copied
+ * here byte for byte, as the MCP server, the agent servers, the Claude Code and
+ * Codex plugins and the VS Code extension copy it. test/claude-md.test.mjs pins
+ * both texts to the builder's sha256, so a change there is a change here, made on
+ * purpose. Never reword them.
+ *
+ * PLAN_USAGE_RULES, the full block, is the kept-current part's "Your plan and its
+ * limits" section, so `pull` brings it to folders an earlier release made.
+ */
+export const PLAN_USAGE_RULES = `Hiveku rules: the account's plan and its limits
+- Every paid plan (Starter, Standard, Premium) opens every feature except the Outbound add-on. Plans differ in how much of each thing they include.
+- Before work that adds something the plan counts (a contact import, a marketing send, phone numbers, social profiles, ad campaigns, websites, automations, databases, AI images or video), call account_usage_get and tell the person what the plan includes, what is used and what happens past the number.
+- A tool answer with limit_kind is not a fault: nothing was done. Never retry it unchanged, never look for a way around it, and never report it as a Hiveku problem. Tell the person in one or two calm sentences what waits and why (the answer's message), and do what its hint says.
+- When limit_kind is plan_limit, not_on_plan or ai_credits, also say when it starts again (resets_at, if any), and the next plan with its number and price (next_plan, next_limit, plans). An owner or admin can upgrade in Settings > Billing (billing_url). When try_next_plan.available is true, they can instead try the next plan free for 14 days, once (try_next_plan.how says where). For billing, add_on, setup and spend_cap, say only what the message and the hint say.
+- Where past is billed, going past the number costs money (rate, up to ceiling): say what it costs before you go past it.
+- Take every number, price and date from account_usage_get or the answer itself, never from memory. Never change the plan, start a try or buy credits or add-ons: an owner or admin does that (billing_url, try_next_plan.how and features.outbound.how say where).
+- Say this only to the account's own people, never to a customer or caller.`;
+
+/**
+ * The builder's PLAN_LIMIT_LINE (same source and pins as PLAN_USAGE_RULES): one
+ * line for AGENTS.md's "Hiveku itself in your way?" item, right after the
+ * memory_write_refused sentence (codex.mjs), so a tool refused at the plan's
+ * number is told to the person and never retried or reported.
+ */
+export const PLAN_LIMIT_LINE =
+  "A tool answer with limit_kind (a plan's number reached, something the plan does not include, AI " +
+  "credits used up, billing that needs attention, an add-on or phones not set up, a chat's spending " +
+  'limit reached) is not a fault: nothing was done. Never retry it unchanged or report it; tell the ' +
+  'person its message and do what its hint says. Only for plan_limit, not_on_plan and ai_credits, also ' +
+  'say when it starts again and the next plan and its number, using account_usage_get or the answer for ' +
+  'every number, and point an owner or admin to its billing_url.';
+
 export const CLAUDE_MD_END = '<!-- hiveku:memory:end -->';
 
 /** The kept-current part: the markers and the two sections between them. */
@@ -62,8 +99,9 @@ export function claudeMdMemoryRegion() {
   return `${CLAUDE_MD_BEGIN}
 ## Folder layout
 
-\`hiveku-sync pull\` keeps this section and "Memory writes" below current. Where
-anything else in this file disagrees with them, follow these two sections.
+\`hiveku-sync pull\` keeps this section, "Memory writes" and "Your plan and its
+limits" below current. Where anything else in this file disagrees with them,
+follow these three sections.
 
 ${SOURCE_OF_TRUTH_LINE}
 ${WORK_LOG_LINE}
@@ -143,6 +181,10 @@ keeps or removes it), and tell the user that owners and admins edit it on the
 Hiveku dashboard under Account memory.
 
 Then run \`hiveku-sync pull\` to refresh the local copies.
+
+## Your plan and its limits
+
+${PLAN_USAGE_RULES}
 ${CLAUDE_MD_END}`;
 }
 

package/src/codex.mjs

@@ -45,7 +45,7 @@ import { formatKey, readTomlDeclarations } from './toml.mjs';
 import { loadDotenv, safeSlug } from './util.mjs';
 import { ownerFolder } from './knowledge-files.mjs';
 import { SHARED_OWNER, ownerLabel, ownerOrder } from './memory-owner.mjs';
-import { SOURCE_OF_TRUTH_LINE, WORK_LOG_LINE } from './claude-md.mjs';
+import { PLAN_LIMIT_LINE, SOURCE_OF_TRUTH_LINE, WORK_LOG_LINE } from './claude-md.mjs';
 
 /**
  * The FIXED list of high-risk Hiveku tools that Codex asks about before
@@ -212,6 +212,7 @@ ${digest}
 - **Hiveku itself in your way?** A Hiveku tool still failing after one sensible retry → \`hiveku_report_issue\`;
   a capability no tool offers (search \`hiveku_docs_search\` first) → \`hiveku_request_feature\`. No secrets or personal data.
   A memory write refused with \`memory_write_refused\` is not a failure: show the person its message and link, follow its hint, and never retry it unchanged or report it.
+  ${PLAN_LIMIT_LINE}
 - **Tell the user only if it changes what they get:** one or two calm sentences — flagged to the Hiveku team (give the
   ref), the team is quick to fix these, you'll let them know when it's sorted. No error codes, blame or promised times.
 - Video generation is paid + capped — \`marketing_generate_video\` with \`dry_run: true\` first.