feat: add /quota slash command for GLM Coding Plan usage display

[william0wang]

Summary

Adds a /quota slash command that queries the GLM Coding Plan usage API directly (no ZCode backend involvement) and renders a multi-line progress-bar card showing remaining usage across windows (5h / weekly / MCP).

• New src/quota/ module (client, parse, format, cache, types, index) — pure, testable units with a single queryQuota() orchestration entry point.
• Slash interception in src/handlers/slash.ts — /quota is the exception: it never calls ZCode, always returns a status line (success card or an error fallback), and never throws.
• Listed in SLASH_COMMANDS so it appears in the editor's / completion menu.

Behavior

• Cached for 10s to debounce rapid repeats; cache is in-memory, process-wide.
• Degrades gracefully: missing apiKey / network failure → unavailable; 401/auth text → auth_error; 429/rate-limit text → rate_limited.
• Host routing mirrors the backend: api.z.ai baseURL → intl endpoint, else CN (open.bigmodel.cn). Credentials reuse the active provider's apiKey from ~/.zcode/v2/config.json.
• Percentage computation has a tolerant fallback chain (remaining+currentValue → currentValue/usage → legacy percentage), so inconsistent field sets across limit kinds still render.

Tests

• tests/quota.test.ts — 26 tests covering parsing (percentage fallbacks, label derivation, error states), formatting (bar rendering, detail sub-lines), TTL cache, and queryQuota orchestration (cache hit / fetch failure → unavailable) with injected fakes (no real network).
• Full suite: 126 tests passing, typecheck clean, 0 lint errors, build OK.

Checklist

• pnpm typecheck
• pnpm lint (0 errors)
• pnpm test
• pnpm build

[gemini-code-assist]

Code Review

This pull request introduces a new /quota slash command that queries the GLM Coding Plan usage API directly and displays the results as a formatted plain-text card. It includes an in-memory TTL cache to debounce requests, an HTTP client, robust response parsing with percentage fallbacks, and comprehensive unit tests. The review feedback suggests several improvements to robustness and type safety: defensively handling missing or malformed properties in item.detail to prevent runtime crashes, explicitly handling HTTP 401 and 403 status codes to correctly return authentication errors, and avoiding non-null assertions (!) when checking payload data.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

Defensively handle potential missing or malformed properties in item.detail elements (such as d.modelCode or d.usage) to prevent runtime TypeError crashes if the API response structure changes or contains unexpected values.

  if (item.detail && item.detail.length > 0) {
    const last = item.detail.length - 1;
    item.detail.forEach((d, i) => {
      if (!d) return;
      const branch = i === last ? "└" : "├";
      const modelCode = typeof d.modelCode === "string" ? d.modelCode : "unknown";
      const usage = typeof d.usage === "number" ? d.usage : 0;
      const name = modelCode.padEnd(DETAIL_LABEL_WIDTH);
      lines.push(`  ${branch} ${name}${usage}`);
    });
  }

[gemini-code-assist]

Explicitly handle HTTP 401 (Unauthorized) and 403 (Forbidden) status codes at the top of parseQuotaEnvelope. If the API gateway returns these status codes with an empty or non-JSON body, the current implementation would fall through and return unavailable instead of auth_error.

  // HTTP-level rate limit or auth errors.
  if (envelope.status === 429 || isRateLimitedMessage(envelope.text)) {
    return { kind: "rate_limited" };
  }
  if (envelope.status === 401 || envelope.status === 403) {
    return { kind: "auth_error" };
  }

[gemini-code-assist]

Avoid using non-null assertion operators (!) by safely checking if payload.data and payload.data.limits exist. This improves type safety and readability.

Suggested change

	const limits = Array.isArray(payload.data?.limits) ? payload.data!.limits! : [];
	const limits = payload.data && Array.isArray(payload.data.limits) ? payload.data.limits : [];