ccqa 1.47.0 → 1.47.2

package/dist/bin/ccqa.mjs

@@ -13535,7 +13535,27 @@ const STEP_EVIDENCE_BEFORE = "ccqaStepBefore";
 const STEP_EVIDENCE_AFTER = "ccqaStepAfter";
 /** The exact boundary call for one step, as emitted and as the gate greps for it. */
 function stepEvidenceCall(fn, marker) {
-	return `await ${fn}(page, ${j(marker.stepId)}, ${j(marker.source)});`;
+	return injectedCall(fn, [j(marker.stepId), j(marker.source)]);
+}
+/**
+* A receiver expression: anything up to the argument's comma, allowing one
+* level of parentheses so `await ctx.newPage()` and `page.context().pages()[1]`
+* — how a rewrite names a tab a click opened — count as receivers.
+*/
+const RECEIVER = String.raw`(?:[^,()]|\([^()]*\))+`;
+/**
+* The call as emitted, plus the pattern that accepts it back on any page. The
+* receiver is the one part a rewrite is right to change: a click that opens a
+* new tab has to act on that tab. Everything else is required verbatim, the
+* `await` and `;` included — an unawaited capture races the end of the test,
+* and a mention in a comment is not a call.
+*/
+function injectedCall(name, args) {
+	const tail = args.map(escapeRegExp).join(String.raw`\s*,\s*`);
+	return {
+		code: `await ${name}(page, ${args.join(", ")});`,
+		pattern: new RegExp(String.raw`await\s+${name}\s*\(\s*${RECEIVER}\s*,\s*${tail}\s*,?\s*\)\s*;`)
+	};
 }
 /**
 * The "preserve the step-evidence calls" rule the library-rewrite prompt must
@@ -13545,7 +13565,7 @@ function stepEvidenceCall(fn, marker) {
 * engine, and the prompt then omits the rule entirely.
 */
 function stepEvidencePreserveRule() {
-	return `**Keep the \`${STEP_EVIDENCE_MODULE}\` calls.** The draft's \`await ${STEP_EVIDENCE_BEFORE}(page, ...)\` / \`await ${STEP_EVIDENCE_AFTER}(page, ...)\` lines are load-bearing: ccqa run reads the per-step screenshots they capture. Keep both calls for every step, in place around that step's actions, with their exact \`(page, "<stepId>", "<source>")\` arguments — and keep the import. If you move a step's actions into a page-object method, leave these two calls in the test body around the call to that method; do NOT move them inside the page object. Never wrap a step in a closure to hold them.`;
+	return `**Keep the \`${STEP_EVIDENCE_MODULE}\` calls.** The draft's \`await ${STEP_EVIDENCE_BEFORE}(page, ...)\` / \`await ${STEP_EVIDENCE_AFTER}(page, ...)\` lines are load-bearing: ccqa run reads the per-step screenshots they capture. Keep both calls for every step, in place around that step's actions, with their exact \`"<stepId>", "<source>"\` arguments — and keep the import. Pass a different page only when the step genuinely acts on one (a click that opens a new tab). If you move a step's actions into a page-object method, leave these two calls in the test body around the call to that method; do NOT move them inside the page object. Never wrap a step in a closure to hold them.`;
 }
 /**
 * Told to the rewrite, because the alternative failure is silent: a claim
@@ -13564,22 +13584,22 @@ function emitPlaywrightDraft(input) {
 	const flushJudgements = (afterActionIndex) => {
 		for (const { step } of judgements.filter((j) => j.afterActionIndex === afterActionIndex)) {
 			if (openMarker) {
-				lines.push(stepEvidenceCall(STEP_EVIDENCE_AFTER, openMarker));
+				lines.push(stepEvidenceCall(STEP_EVIDENCE_AFTER, openMarker).code);
 				openMarker = null;
 			}
 			if (lines.length > 0) lines.push("");
 			lines.push(`// step: ${step.id} [${step.source}]`);
-			lines.push(judgeCall(step));
+			lines.push(judgeCall(step).code);
 		}
 	};
 	flushJudgements(-1);
 	for (let i = 0; i < actions.length; i++) {
 		const marker = markerByIndex.get(i);
 		if (marker) {
-			if (openMarker) lines.push(stepEvidenceCall(STEP_EVIDENCE_AFTER, openMarker));
+			if (openMarker) lines.push(stepEvidenceCall(STEP_EVIDENCE_AFTER, openMarker).code);
 			if (lines.length > 0) lines.push("");
 			lines.push(`// step: ${marker.stepId} [${marker.source}]`);
-			lines.push(stepEvidenceCall(STEP_EVIDENCE_BEFORE, marker));
+			lines.push(stepEvidenceCall(STEP_EVIDENCE_BEFORE, marker).code);
 			openMarker = marker;
 		}
 		const action = actions[i];
@@ -13591,7 +13611,7 @@ function emitPlaywrightDraft(input) {
 		}
 		flushJudgements(i);
 	}
-	if (openMarker) lines.push(stepEvidenceCall(STEP_EVIDENCE_AFTER, openMarker));
+	if (openMarker) lines.push(stepEvidenceCall(STEP_EVIDENCE_AFTER, openMarker).code);
 	if (judgements.length > 0) lines.unshift("test.slow();", "");
 	const body = lines.map((l) => l === "" ? "" : `  ${l}`).join("\n");
 	return [
@@ -13779,8 +13799,9 @@ const j = (s) => JSON.stringify(s);
 const jExpr = (s) => envRefsToJsExpression(s);
 /** One claim, asserted through the judge. Exported so the generation gate can require it back. */
 function judgeCall(step) {
-	const from = step.from === void 0 ? "" : `, ${jExpr(step.from)}`;
-	return `await ${JUDGE_CALL}(page, ${bracedRefsToJsExpression(step.judgeByLlm.trim())}${from});`;
+	const args = [bracedRefsToJsExpression(step.judgeByLlm.trim())];
+	if (step.from !== void 0) args.push(jExpr(step.from));
+	return injectedCall(JUDGE_CALL, args);
 }
 //#endregion
 //#region src/targets/playwright/browser-server.ts
@@ -14031,12 +14052,12 @@ async function missingInjectedCalls(result, markers, judgements) {
 	const corpus = (await Promise.all(result.files.filter((f) => f.kind === "test").map((f) => readFile(f.path, "utf8").catch(() => "")))).join("\n");
 	const warnings = [];
 	for (const m of markers) {
-		const hasBefore = corpus.includes(stepEvidenceCall(STEP_EVIDENCE_BEFORE, m));
-		const hasAfter = corpus.includes(stepEvidenceCall(STEP_EVIDENCE_AFTER, m));
+		const hasBefore = stepEvidenceCall(STEP_EVIDENCE_BEFORE, m).pattern.test(corpus);
+		const hasAfter = stepEvidenceCall(STEP_EVIDENCE_AFTER, m).pattern.test(corpus);
 		if (!hasBefore || !hasAfter) warnings.push(`step ${m.stepId}: generated test is missing its ${STEP_EVIDENCE_BEFORE}/${STEP_EVIDENCE_AFTER} call(s) — that step will have no report screenshots. A rewrite pass must not drop them.`);
 	}
 	for (const { step } of judgements) {
-		if (corpus.includes(judgeCall(step))) continue;
+		if (judgeCall(step).pattern.test(corpus)) continue;
 		warnings.push(`step ${step.id}: generated test is missing its ${JUDGE_CALL} call — that claim is never decided and the spec passes without testing it. A rewrite pass must not drop or reword it.`);
 	}
 	return warnings;
@@ -16431,6 +16452,7 @@ find nth <index> "<ALLOWED-css>" <action>
 5. \`--name "<n>"\` is **role-only**. Never pass it to \`find text\`, \`find label\`, etc.
 6. \`--name\` matches by substring. Add \`--exact\` whenever the name you read from the snapshot is the element's whole accessible name — without it a short name such as "Log in" also matches "Log in with Google" and "Log in with SSO", and the recording silently pins the first of them.
 7. \`find\` includes its own wait; do not chain a \`wait\` before it.
+8. **A step that names an element by position gets a positional locator.** "the first row", "the topmost item", "the latest message" identify by \`first\`/\`last\`/\`nth\` on a stable inner selector — never by the title, subject, or body text that record happened to carry. That text is data the environment supplies: pinning it makes the test pass only while that same record stays in that position, and the next run opens a different one.
 
 **Examples:**
 
@@ -16611,6 +16633,8 @@ CCQA_STEP=<step-id> CCQA_ASSERT=url_contains:/dashboard agent-browser --session
 - A marked command that exits non-zero records nothing — fix the check and re-run it.
 - A marker that doesn't match its command (e.g. \`CCQA_ASSERT=1\` on a \`click\`) is ignored with a warning — never do that.
 - \`get count\` and \`get url\` exit 0 regardless of what they print. **Read the output**: if the printed count / URL contradicts the marker, the signal did NOT verify — treat it as a failed verification (emit \`ASSERTION_FAILED\` if the step cannot be confirmed another way).
+- **Assert the thing, not how much of it there is.** A step that says a list, a table or a section is shown is satisfied by the list being there. Do not assert the count it happened to have, and never assert an empty-state message ("no items yet") unless the step's own \`expected\` asks for emptiness — the next run has one more record than this one and the test breaks for a reason nobody chose.
+
 - **Assert what the step asks about, nothing else.** The \`expected\` is the contract; anything else you happened to see on the way is not. A nav item, a heading or a greeting that the step never mentions adds no coverage, differs between recordings of the same spec, and is the first thing to break on replay — so the next recording quietly drops it and the test gets weaker without anyone deciding that.
 
 - **\`url_contains\` is opt-in, not a habit.** The same rule, in the form that gets broken most. Emit it ONLY when a step's own \`expected\` explicitly asks about the URL or path. Do NOT add a \`url_contains\` to "prove" a login succeeded, a page loaded, or a navigation happened — confirm those with \`text_visible\` / \`element_visible\` on something the destination page renders. An unrequested URL assertion adds no coverage the visible-content assert doesn't already give, and is the single most common way an environment gets baked into a test.

package/dist/hub-client/index.d.mts

@@ -35,9 +35,9 @@ declare const RunSchema: z.ZodObject<{
     running: "running";
   }>;
   kind: z.ZodDefault<z.ZodEnum<{
+    record: "record";
     run: "run";
     drift: "drift";
-    record: "record";
   }>>;
   drift: z.ZodDefault<z.ZodNullable<z.ZodObject<{
     specs: z.ZodNumber;
@@ -811,9 +811,9 @@ type ReportSpecResult = z.infer<typeof ReportSpecResultSchema>;
 declare const RunReportDataSchema: z.ZodObject<{
   schemaVersion: z.ZodLiteral<1>;
   kind: z.ZodDefault<z.ZodEnum<{
+    record: "record";
     run: "run";
     drift: "drift";
-    record: "record";
   }>>;
   createdAt: z.ZodString;
   runId: z.ZodNullable<z.ZodString>;

package/dist/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ccqa",
-  "version": "1.47.0",
+  "version": "1.47.2",
   "type": "module",
   "description": "Browser test recorder powered by Claude Code and agent-browser",
   "repository": {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ccqa",
-  "version": "1.47.0",
+  "version": "1.47.2",
   "type": "module",
   "description": "Browser test recorder powered by Claude Code and agent-browser",
   "repository": {