@tokenzero/mcp 0.2.6 → 0.2.8

package/README.md

@@ -15,6 +15,8 @@ claude mcp add tokenzero -- npx -y @tokenzero/mcp     # Claude Code
 | `audit_repo_token_waste` | locally | Static audit of a local directory (default: the working directory), a public GitHub repo, or a diff. Nothing from a local directory is uploaded. |
 | `calculate_model_arbitrage` | locally | Prices a workload across models at list prices, with and without prompt caching. |
 | `optimize_prompt_payload` | TokenZero service | Rewrites a request body so its static prefix is cacheable. Sends the body you pass to the TokenZero service (`https://tokenzero.sh` unless `TOKENZERO_URL` is set); do not include secrets. |
+| `workspace_metrics` | TokenZero service | Reads a workspace's savings scorecard: hard savings and net ROI per window, each engine's contribution and one-click recommendations. Uses `TOKENZERO_KEY` unless a key is passed; `demo: true` returns a labelled sample. |
+| `provision_workspace` | TokenZero service | Creates a free workspace and key with no form. The key is returned once. |
 
 Optional environment: `TOKENZERO_URL` (another gateway for `optimize_prompt_payload`; default `https://tokenzero.sh`), `GITHUB_TOKEN` (raises GitHub rate limits for repo audits).
 

package/bin/tokenzero-mcp.mjs

@@ -9,17 +9,21 @@ import readline from "node:readline";
 import crypto from "node:crypto";
 
 // src/pricing.ts
-var PRICING_VERSION = "2026-10-26";
+var PRICING_VERSION = "2026-10-27";
 var P = (provider, tier, input, output, cachedInput, cacheWrite, extra = {}) => ({ provider, tier, input, output, cachedInput, cacheWrite, ...extra });
 var PRICES = {
   // ---- OpenAI (standard tier, ≤272K; cache write = 30m TTL rate) ----
   "gpt-6-astra": P("openai", "frontier", 10, 50, 1, 12.5, { ultrafast: true, longCtxAt: 272e3, longCtx: { input: 20, output: 75, cachedInput: 2, cacheWrite: 25 }, minCacheTokens: 1024 }),
   "gpt-6.1-sol": P("openai", "mid", 2, 10, 0.1, 2.5, { ultrafast: true, longCtxAt: 272e3, longCtx: { input: 4, output: 15, cachedInput: 0.2, cacheWrite: 5 }, minCacheTokens: 1024 }),
+  "gpt-6.1-terra": P("openai", "mid", 1, 5, 0.05, 1.25, { longCtxAt: 272e3, longCtx: { input: 2, output: 7.5, cachedInput: 0.1, cacheWrite: 2.5 }, minCacheTokens: 1024 }),
   "gpt-6-sol": P("openai", "mid", 2, 10, 0.2, 2.5, { longCtxAt: 272e3, longCtx: { input: 4, output: 15, cachedInput: 0.4, cacheWrite: 5 }, minCacheTokens: 1024 }),
   "gpt-6-luna": P("openai", "fast", 0.1, 0.5, 0.01, 0.125, { longCtxAt: 272e3, longCtx: { input: 0.2, output: 0.75, cachedInput: 0.02, cacheWrite: 0.25 }, minCacheTokens: 1024 }),
   "gpt-5.6-sol": P("openai", "frontier", 4, 20, 0.4, 5, { longCtxAt: 272e3, longCtx: { input: 8, output: 30, cachedInput: 0.8, cacheWrite: 10 }, minCacheTokens: 1024 }),
   "gpt-5.6-terra": P("openai", "mid", 2, 12, 0.2, 2.5, { longCtxAt: 272e3, longCtx: { input: 4, output: 18, cachedInput: 0.4, cacheWrite: 5 }, minCacheTokens: 1024 }),
   "gpt-5.6-luna": P("openai", "fast", 0.2, 1.2, 0.02, 0.25, { longCtxAt: 272e3, longCtx: { input: 0.4, output: 1.8, cachedInput: 0.04, cacheWrite: 0.5 }, minCacheTokens: 1024 }),
+  // Daybreak cyber model: access needs OpenAI approval; 400K window
+  "gpt-5.6-cyber": P("openai", "frontier", 12.5, 75, 1.25, 15.625, { longCtxAt: 272e3, longCtx: { input: 25, output: 112.5, cachedInput: 2.5, cacheWrite: 31.25 }, minCacheTokens: 1024 }),
+  "gpt-5.3-codex": P("openai", "mid", 1.75, 14, 0.175, void 0, { minCacheTokens: 1024 }),
   "gpt-5.5": P("openai", "frontier", 5, 30, 0.5, void 0, { fastFactor: 2.5, longCtxAt: 272e3, longCtx: { input: 10, output: 45, cachedInput: 1 }, minCacheTokens: 1024 }),
   "gpt-5.5-pro": P("openai", "frontier", 30, 180),
   "gpt-5.4": P("openai", "frontier", 2.5, 15, 0.25, void 0, { longCtxAt: 272e3, longCtx: { input: 5, output: 22.5, cachedInput: 0.5 }, minCacheTokens: 1024 }),
@@ -84,7 +88,8 @@ var PRICES = {
 var MODEL_ALIASES = { "deepseek-v4-flash": "deepseek-flash", "deepseek-v4-flash-vision-exp": "deepseek-flash" };
 var CASCADE_TARGET = {
   "gpt-6-astra": "gpt-6.1-sol",
-  "gpt-6.1-sol": "gpt-6-luna",
+  "gpt-6.1-sol": "gpt-6.1-terra",
+  "gpt-6.1-terra": "gpt-6-luna",
   "gpt-6-sol": "gpt-6-luna",
   "gpt-5.6-sol": "gpt-5.6-terra",
   "gpt-5.6-terra": "gpt-5.6-luna",
@@ -157,6 +162,7 @@ var MODEL_SUNSET = {
   "gpt-5-nano": { retires: "2026-12-11", replacement: "gpt-5.6-luna", announced: "2026-06-11" },
   "gpt-5-pro": { retires: "2026-12-11", replacement: "gpt-5.6-sol", announced: "2026-06-11" },
   "gpt-5.1": { retires: "2027-04-01", replacement: "gpt-6-sol", announced: "2026-10-01" },
+  "gpt-5.3-codex": { retires: "2027-04-01", replacement: "gpt-6-sol", announced: "2026-10-01" },
   "gpt-5.4-nano": { retires: "2027-04-01", replacement: "gpt-6-luna", announced: "2026-10-01" },
   "o1": { retires: "2026-10-23", replacement: "gpt-5.6-sol", announced: "2026-04-22" },
   "o3": { retires: "2026-12-11", replacement: "gpt-5.6-sol", announced: "2026-06-11" },
@@ -637,7 +643,7 @@ async function fetchRepoFiles(repo, opts = {}) {
   const meta = await fetch(`${GH}/repos/${repo}`, { headers: ghHeaders(opts.token) });
   if (meta.status === 404) throw new Error("repository not found (private repos are not supported)");
   if (!meta.ok) throw new Error(`GitHub API ${meta.status}: ${(await meta.text()).slice(0, 200)}`);
-  const branch = (await meta.json()).default_branch;
+  const metaJson = await meta.json(), branch = metaJson.default_branch, fullName = typeof metaJson.full_name === "string" ? metaJson.full_name : repo;
   const tree = await fetch(`${GH}/repos/${repo}/git/trees/${encodeURIComponent(branch)}?recursive=1`, { headers: ghHeaders(opts.token) });
   if (!tree.ok) throw new Error(`GitHub tree ${tree.status}`);
   const treeAll = (await tree.json()).tree;
@@ -661,7 +667,7 @@ async function fetchRepoFiles(repo, opts = {}) {
     const r = await fetch(`https://raw.githubusercontent.com/${repo}/${branch}/${n}`, { headers: { "user-agent": UA } });
     return r.ok ? { path: n, content: await r.text() } : null;
   }))).filter((m) => !!m);
-  return { files, manifests: manifests2, branch, total_candidates: entries.length };
+  return { files, manifests: manifests2, branch, total_candidates: entries.length, full_name: fullName };
 }
 
 // src/arbitrage.ts
@@ -754,7 +760,7 @@ function manifests(root) {
 }
 
 // mcp/src/main.ts
-var VERSION = "0.2.6";
+var VERSION = "0.2.8";
 var BAKED_URL = true ? "https://tokenzero.sh" : "__TOKENZERO_DEFAULT_URL__";
 var SUPPORTED = [...MCP_LEGACY_VERSIONS];
 var SERVER = { name: "tokenzero", version: VERSION };
@@ -764,26 +770,36 @@ var serviceUrl = () => (process.env.TOKENZERO_URL || (BAKED_URL.startsWith("http
 var TOOLS = [
   {
     name: "audit_repo_token_waste",
+    title: "Audit a repo or directory for LLM token waste",
+    annotations: { title: "Audit a repo or directory for LLM token waste", readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: true },
     description: "Static audit for LLM token waste: uncached static prompts and prompt templates, frontier models on extraction/classification prompts, calls in loops, unpriced models. Give `path` for a local directory (scanned on this machine, nothing uploaded; defaults to the working directory), `repo` (owner/name) for a public GitHub repo, or `diff` for a unified diff. Results are static heuristics, not measurements.",
     inputSchema: { type: "object", properties: { path: { type: "string", description: "local directory; defaults to the current working directory" }, repo: { type: "string", description: "public GitHub owner/name" }, diff: { type: "string", description: "unified diff text, e.g. from git diff" } }, additionalProperties: false }
   },
   {
     name: "calculate_model_arbitrage",
+    title: "Price a workload across models",
+    annotations: { title: "Price a workload across models", readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: false },
     description: `Price a workload at list prices (table ${PRICING_VERSION}) for the requested model, its cheaper sibling and any extra models, with and without prompt caching. Runs locally.`,
     inputSchema: { type: "object", properties: { model: { type: "string" }, input_tokens: { type: "number" }, output_tokens: { type: "number" }, requests_per_month: { type: "number" }, static_tokens: { type: "number", description: "portion of input_tokens that is a fixed prefix (system + tools)" }, compare_to: { type: "array", items: { type: "string" } } }, required: ["model", "input_tokens", "output_tokens"], additionalProperties: false }
   },
   {
     name: "optimize_prompt_payload",
+    title: "Make a request body cache-friendly",
+    annotations: { title: "Make a request body cache-friendly", readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: false },
     description: "Rewrite an OpenAI Chat/Responses or Anthropic Messages request body so its static prefix is provider-cacheable, and project the saving. Sends the body you pass to the TokenZero service (set TOKENZERO_URL); do not pass secrets in it.",
     inputSchema: { type: "object", properties: { body: { type: "object" }, endpoint: { type: "string", enum: ["/v1/chat/completions", "/v1/responses", "/v1/messages"] }, strategy: { type: "string", enum: ["static", "conversation"] } }, required: ["body"], additionalProperties: false }
   },
   {
     name: "workspace_metrics",
+    title: "Read a workspace savings scorecard",
+    annotations: { title: "Read a workspace savings scorecard", readOnlyHint: true, destructiveHint: false, idempotentHint: true, openWorldHint: false },
     description: "Live value scorecard for a TokenZero workspace (GET /v1/workspaces/me/metrics): hard savings and net ROI per window, per-engine contribution, speed, parity, availability, privacy posture, per-model breakdown and one-click recommendations. Uses TOKENZERO_KEY unless a key is given; demo: true returns a labelled sample.",
     inputSchema: { type: "object", properties: { key: { type: "string", description: "workspace key (tz_live_\u2026); defaults to TOKENZERO_KEY" }, window: { type: "string", enum: ["24h", "7d", "30d", "all"] }, demo: { type: "boolean" } } }
   },
   {
     name: "provision_workspace",
+    title: "Create a free TokenZero workspace",
+    annotations: { title: "Create a free TokenZero workspace", readOnlyHint: false, destructiveHint: false, idempotentHint: false, openWorldHint: false },
     description: "Create a TokenZero workspace and key with no form (POST /v1/workspaces on the TokenZero service). Returns the key once, the base URLs that route through it, the one-click dashboard link (key in the URL fragment, never sent to the server) and the badge URL. Free up to the monthly free tier; nothing is billed.",
     inputSchema: { type: "object", properties: { name: { type: "string", description: 'workspace name (defaults to "mcp-agent")' } }, additionalProperties: false }
   }

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@tokenzero/mcp",
   "mcpName": "sh.tokenzero/mcp",
-  "version": "0.2.6",
+  "version": "0.2.8",
   "description": "TokenZero MCP server (stdio): audit LLM token waste in your project locally, price model alternatives, and make prompts cache-friendly from Claude Code, Cursor, Windsurf or Cline.",
   "type": "module",
   "license": "Apache-2.0",