docs(ci): add coverage buffer and contributor test-quality guidance

.github/ISSUE_TEMPLATE/feature_request.yml

@@ -49,6 +49,13 @@ body:
         - ...
     validations:
       required: true
+  - type: textarea
+    id: test_scenarios
+    attributes:
+      label: Test scenarios
+      description: List unit/integration cases for new branches, fallback paths, sanitizer rules, and regressions. Call out public/private safety boundaries if GitHub comments or MCP output changes.
+    validations:
+      required: true
   - type: checkboxes
     id: boundaries
     attributes:
@@ -62,3 +69,5 @@ body:
           required: true
         - label: This does not auto-close, auto-merge, rewrite contributor work, or label PRs outside the confirmed-miner policy.
           required: true
+        - label: Public GitHub output will be tested against forbidden language (wallet, hotkey, raw trust score, payout, reward estimate, farming, private reviewability, public score estimate).
+          required: true

.github/pull_request_template.md

@@ -13,7 +13,7 @@
 - [ ] `git diff --check`
 - [ ] `npm run actionlint`
 - [ ] `npm run typecheck`
-- [ ] `npm run test:coverage`
+- [ ] `npm run test:coverage` locally; global coverage stays at or above **97%** for lines, statements, functions, and branches (aim for **98%+** branch coverage locally so CI variance does not fail near the threshold)
 - [ ] `npm run test:workers`
 - [ ] `npm run build:mcp`
 - [ ] `npm run test:mcp-pack`
@@ -22,7 +22,7 @@
 - [ ] `npm run ui:typecheck`
 - [ ] `npm run ui:build`
 - [ ] `npm audit --audit-level=moderate`
-- [ ] Coverage remains at or above 97% for statements, branches, functions, and lines.
+- [ ] New or changed behavior has unit/integration tests for new branches, fallback paths, and sanitizer boundaries
 
 If any required check was skipped, explain why:
 

CONTRIBUTING.md

@@ -86,8 +86,10 @@ npm run ui:build
 npm audit --audit-level=moderate
 ```
 
-`npm run test:ci` runs the normal combined gate. Coverage must stay at or above 97% for
-statements, branches, functions, and lines.
+`npm run test:ci` runs the normal combined gate. Coverage must stay at or above **97%** for
+statements, branches, functions, and lines. Run `npm run test:coverage` locally when you change
+behavior, and aim for **98%+ branch coverage** locally so small CI variance does not fail near the
+threshold.
 
 Maintainer/release smoke checks:
 
@@ -103,6 +105,13 @@ The browser smoke path is manual until it is stable enough to make required on e
 
 Tests should prove behavior, not just exercise lines for coverage.
 
+- Add or update tests for every behavior change: new branches, fallback paths, sanitizer rules,
+  and regressions.
+- Add invariant or property-style tests when behavior depends on sorting, gating, public/private
+  boundaries, scoring, queue pressure, or source-upload safety.
+- Public GitHub comments must be tested against forbidden language when comment text changes
+  (wallet, hotkey, raw trust score, payout, reward estimate, farming, private reviewability,
+  public score estimate).
 - Backend/API tests should cover success, denied, invalid input, missing auth, scoped auth,
   rate-limit, persistence, and error-shaping paths.
 - Auth tests should cover browser cookie sessions, bearer sessions, logout/revocation, GitHub
@@ -146,6 +155,8 @@ MCP:
 
 - Preserve backwards compatibility where practical.
 - Keep CLI output stable and JSON output parseable.
+- `gittensory-mcp init-client --print` supports `codex`, `claude`, `cursor`, and `mcp` (generic
+  JSON hosts that use the `mcpServers` shape).
 - MCP package releases are prepared separately and published from protected `mcp-vX.Y.Z` tags.
 
 Public GitHub surfaces:

apps/gittensory-ui/src/routes/docs.mcp-clients.tsx

@@ -32,6 +32,20 @@ function McpClients() {
       title="MCP client setup"
       description="Configure your coding agent to talk to the Gittensory MCP. Pick stdio for local agents, remote for cloud agents."
     >
+      <h2>Generate config</h2>
+      <p>These commands print config only. They do not mutate your local client files.</p>
+      <CodeBlock
+        lang="bash"
+        code={`gittensory-mcp init-client --print codex
+gittensory-mcp init-client --print claude
+gittensory-mcp init-client --print cursor
+gittensory-mcp init-client --print mcp`}
+      />
+      <p>
+        <code>--print mcp</code> uses the same JSON snippet as Claude Desktop and Cursor for other
+        stdio MCP hosts that expect the <code>mcpServers</code> shape.
+      </p>
+
       <h2>Codex (OpenAI)</h2>
       <CodeBlock
         filename="~/.codex/config.toml"

packages/gittensory-mcp/bin/gittensory-mcp.js

@@ -593,7 +593,7 @@ function printHelp() {
   gittensory-mcp status [--json]
   gittensory-mcp changelog [--json]
   gittensory-mcp doctor [--cwd path] [--json]
-  gittensory-mcp init-client --print codex|claude|cursor [--json]
+  gittensory-mcp init-client --print codex|claude|cursor|mcp [--json]
   gittensory-mcp analyze-branch --login <github-login> [--repo owner/repo] [--base origin/main] [--pending-merged-prs 3] [--expected-open-prs 0] [--projected-credibility 0.8] [--scenario-note "..."] [--validation "passed|npm test|summary"] [--json]
   gittensory-mcp preflight --login <github-login> [--repo owner/repo] [--base origin/main] [--pending-merged-prs 3] [--expected-open-prs 0] [--projected-credibility 0.8] [--validation "passed|npm test|summary"] [--json]
   gittensory-mcp agent plan --login <github-login> [--repo owner/repo] [--json]
@@ -852,7 +852,7 @@ async function doctor(options) {
 
   const commandPath = findExecutable("gittensory-mcp");
   if (commandPath) add("client_path", "pass", "gittensory-mcp is visible on PATH.");
-  else add("client_path", "warn", "gittensory-mcp was not found on PATH.", "Use an absolute command path in Codex, Claude, or MCP client config.");
+  else add("client_path", "warn", "gittensory-mcp was not found on PATH.", "Use an absolute command path in your MCP client config.");
 
   const scorerCommand = resolveScorePreviewCommand();
   if (!scorerCommand) {
@@ -897,7 +897,7 @@ async function doctor(options) {
 
 function initClient(options) {
   const client = String(options.print ?? options.client ?? "").toLowerCase();
-  if (!client) throw new Error("Pass --print codex, --print claude, or --print cursor.");
+  if (!client) throw new Error("Pass --print codex, --print claude, --print cursor, or --print mcp.");
   const command = options.command ?? "gittensory-mcp";
   const snippet = clientSnippet(client, command);
   const payload = {
@@ -1067,7 +1067,7 @@ function redactPrivateValidationMetrics(text) {
 
 function clientSnippet(client, command) {
   if (client === "codex") return `[mcp_servers.gittensory]\ncommand = ${JSON.stringify(command)}\nargs = ["--stdio"]`;
-  if (client === "claude" || client === "cursor") {
+  if (client === "claude" || client === "cursor" || client === "mcp") {
     return JSON.stringify(
       {
         mcpServers: {
@@ -1081,7 +1081,7 @@ function clientSnippet(client, command) {
       2,
     );
   }
-  throw new Error(`Unsupported client: ${client}. Use codex, claude, or cursor.`);
+  throw new Error(`Unsupported client: ${client}. Use codex, claude, cursor, or mcp.`);
 }
 
 function findExecutable(name) {

test/unit/mcp-cli.test.ts

@@ -26,6 +26,12 @@ describe("gittensory-mcp CLI", () => {
     const claude = JSON.parse(run(["init-client", "--print", "claude", "--json"])) as { snippet: string };
     expect(claude.snippet).toContain('"mcpServers"');
     expect(claude.snippet).toContain('"gittensory"');
+
+    const cursor = JSON.parse(run(["init-client", "--print", "cursor", "--json"])) as { snippet: string };
+    expect(cursor.snippet).toBe(claude.snippet);
+
+    const generic = JSON.parse(run(["init-client", "--print", "mcp", "--json"])) as { snippet: string };
+    expect(generic.snippet).toBe(claude.snippet);
   });
 
   it("runs doctor against a local health/session fixture", async () => {