substrate-ai 0.20.39 → 0.20.41

package/dist/cli/index.js

@@ -1,10 +1,10 @@
 #!/usr/bin/env node
-import { FileStateStore, RunManifest, SUBSTRATE_OWNED_SETTINGS_KEYS, SupervisorLock, VALID_PHASES, WorkGraphRepository, ZERO_FINDING_COUNTS, buildPipelineStatusOutput, createDatabaseAdapter, createStateStore, findPackageRoot, formatOutput, formatPipelineStatusHuman, formatPipelineSummary, formatTokenTelemetry, getAllDescendantPids, getAutoHealthData, getSubstrateDefaultSettings, inspectProcessTree, parseDbTimestampAsUtc, parseRuntimeProbes, registerHealthCommand, resolveBmadMethodSrcPath, resolveBmadMethodVersion, resolveMainRepoRoot, resolveRunManifest, rollupFindingCounts } from "../health-R9GiVIiQ.js";
+import { FileStateStore, RunManifest, SUBSTRATE_OWNED_SETTINGS_KEYS, SupervisorLock, VALID_PHASES, WorkGraphRepository, ZERO_FINDINGS_BY_AUTHOR, ZERO_FINDING_COUNTS, ZERO_PROBE_AUTHOR_METRICS, aggregateProbeAuthorMetrics, buildPipelineStatusOutput, createDatabaseAdapter, createStateStore, findPackageRoot, formatOutput, formatPipelineStatusHuman, formatPipelineSummary, formatTokenTelemetry, getAllDescendantPids, getAutoHealthData, getSubstrateDefaultSettings, inspectProcessTree, parseDbTimestampAsUtc, parseRuntimeProbes, registerHealthCommand, resolveBmadMethodSrcPath, resolveBmadMethodVersion, resolveMainRepoRoot, resolveRunManifest, rollupFindingCounts, rollupFindingsByAuthor, rollupProbeAuthorMetrics } from "../health-BmEu3n9Z.js";
 import { createLogger } from "../logger-KeHncl-f.js";
 import { createEventBus } from "../helpers-CElYrONe.js";
 import { AdapterRegistry, BudgetConfigSchema, CURRENT_CONFIG_FORMAT_VERSION, CURRENT_TASK_GRAPH_VERSION, ConfigError, CostTrackerConfigSchema, DEFAULT_CONFIG, DoltClient, DoltNotInstalled, GlobalSettingsSchema, InMemoryDatabaseAdapter, IngestionServer, MonitorDatabaseImpl, OPERATIONAL_FINDING, PartialGlobalSettingsSchema, PartialProviderConfigSchema, ProvidersSchema, RoutingRecommender, STORY_METRICS, TelemetryConfigSchema, addTokenUsage, aggregateTokenUsageForRun, checkDoltInstalled, compareRunMetrics, createAmendmentRun, createConfigSystem, createDecision, createDoltClient, createPipelineRun, getActiveDecisions, getAllCostEntriesFiltered, getBaselineRunMetrics, getDecisionsByCategory, getDecisionsByPhaseForRun, getLatestCompletedRun, getLatestRun, getPipelineRunById, getPlanningCostTotal, getRetryableEscalations, getRunMetrics, getRunningPipelineRuns, getSessionCostSummary, getSessionCostSummaryFiltered, getStoryMetricsForRun, getTokenUsageSummary, incrementRunRestarts, initSchema, initializeDolt, listRunMetrics, loadParentRunDecisions, supersedeDecision, tagRunAsBaseline, updatePipelineRun } from "../dist-VcMmfo2w.js";
 import "../adapter-registry-DXLMTmfD.js";
-import { AdapterTelemetryPersistence, AppError, DoltRepoMapMetaRepository, DoltSymbolRepository, ERR_REPO_MAP_STORAGE_WRITE, EpicIngester, GitClient, GrammarLoader, RepoMapInjector, RepoMapModule, RepoMapQueryEngine, RepoMapStorage, SymbolParser, createContextCompiler, createDispatcher, createEventEmitter, createImplementationOrchestrator, createPackLoader, createPhaseOrchestrator, createStopAfterGate, createTelemetryAdvisor, formatPhaseCompletionSummary, getFactoryRunSummaries, getScenarioResultsForRun, getTwinRunsForRun, listGraphRuns, registerExportCommand, registerFactoryCommand, registerRunCommand, registerScenariosCommand, resolveStoryKeys, runAnalysisPhase, runPlanningPhase, runProbeAuthor, runSolutioningPhase, validateStopAfterFromConflict } from "../run-Dm70k8bR.js";
+import { AdapterTelemetryPersistence, AppError, DoltRepoMapMetaRepository, DoltSymbolRepository, ERR_REPO_MAP_STORAGE_WRITE, EpicIngester, GitClient, GrammarLoader, RepoMapInjector, RepoMapModule, RepoMapQueryEngine, RepoMapStorage, SymbolParser, createContextCompiler, createDispatcher, createEventEmitter, createImplementationOrchestrator, createPackLoader, createPhaseOrchestrator, createStopAfterGate, createTelemetryAdvisor, formatPhaseCompletionSummary, getFactoryRunSummaries, getScenarioResultsForRun, getTwinRunsForRun, listGraphRuns, registerExportCommand, registerFactoryCommand, registerRunCommand, registerScenariosCommand, resolveStoryKeys, runAnalysisPhase, runPlanningPhase, runProbeAuthor, runSolutioningPhase, validateStopAfterFromConflict } from "../run-DnwsCYYA.js";
 import "../errors-CogpxBUg.js";
 import "../routing-CcBOCuC9.js";
 import "../decisions-C0pz9Clx.js";
@@ -3667,7 +3667,7 @@ async function runStatusAction(options) {
 			logger$13.debug({ err }, "Work graph query failed, continuing without work graph data");
 		}
 		if (run === void 0) {
-			const { inspectProcessTree: inspectProcessTree$1 } = await import("../health-C-ZrHJwL.js");
+			const { inspectProcessTree: inspectProcessTree$1 } = await import("../health-CAV-PNE4.js");
 			const substrateDirPath = join(projectRoot, ".substrate");
 			const processInfo = inspectProcessTree$1({
 				projectRoot,
@@ -3712,6 +3712,8 @@ async function runStatusAction(options) {
 				} catch {}
 				const verificationResult = manifestPerStoryState?.[row.story_key]?.verification_result;
 				const verificationFindings = rollupFindingCounts(verificationResult);
+				const probeAuthorMetrics = rollupProbeAuthorMetrics(verificationResult);
+				const findingsByAuthor = rollupFindingsByAuthor(verificationResult);
 				const verificationRan = verificationResult !== void 0 && verificationResult !== null;
 				return {
 					story_key: row.story_key,
@@ -3724,8 +3726,12 @@ async function runStatusAction(options) {
 					},
 					review_cycles: row.review_cycles ?? 0,
 					dispatches: row.dispatches ?? 0,
-					verification_findings: verificationFindings,
-					verification_ran: verificationRan
+					verification_findings: {
+						...verificationFindings,
+						byAuthor: findingsByAuthor
+					},
+					verification_ran: verificationRan,
+					probe_author: probeAuthorMetrics
 				};
 			});
 			let pipelineWallClockMs = 0;
@@ -5198,7 +5204,7 @@ async function runSupervisorAction(options, deps = {}) {
 								await initSchema(expAdapter);
 								const { runRunAction: runPipeline } = await import(
 									/* @vite-ignore */
-									"../run-CL1C4M2f.js"
+									"../run-CXTAvjCK.js"
 );
 								const runStoryFn = async (opts) => {
 									const exitCode = await runPipeline({
@@ -5552,7 +5558,7 @@ function printFactoryRunTable(runs) {
 	}
 }
 async function runMetricsAction(options) {
-	const { outputFormat, projectRoot, limit = 10, compare, tagBaseline, analysis, sprint, story, taskType, since, aggregate, efficiency, recommendations, turns, consumers, categories, compareStories, routingRecommendations, run, factory } = options;
+	const { outputFormat, projectRoot, limit = 10, compare, tagBaseline, analysis, sprint, story, taskType, since, aggregate, efficiency, recommendations, turns, consumers, categories, compareStories, routingRecommendations, run, factory, probeAuthorSummary } = options;
 	const telemetryModes = [
 		efficiency,
 		recommendations,
@@ -5962,6 +5968,8 @@ async function runMetricsAction(options) {
 		} catch {}
 		const findingCountsByStoryRun = new Map();
 		const verificationRanByStoryRun = new Map();
+		const probeAuthorByStoryRun = new Map();
+		const findingsByAuthorByStoryRun = new Map();
 		const uniqueRunIds = Array.from(new Set(storyMetrics.map((sm) => sm.run_id).filter((id) => id !== "")));
 		for (const uniqueRunId of uniqueRunIds) try {
 			const { manifest } = await resolveRunManifest(dbRoot, uniqueRunId);
@@ -5971,6 +5979,8 @@ async function runMetricsAction(options) {
 				const key = `${storyKey}:${uniqueRunId}`;
 				findingCountsByStoryRun.set(key, rollupFindingCounts(entry.verification_result));
 				verificationRanByStoryRun.set(key, entry.verification_result !== void 0 && entry.verification_result !== null);
+				probeAuthorByStoryRun.set(key, rollupProbeAuthorMetrics(entry.verification_result));
+				findingsByAuthorByStoryRun.set(key, rollupFindingsByAuthor(entry.verification_result));
 			}
 		} catch {}
 		let factoryRuns = [];
@@ -5985,16 +5995,30 @@ async function runMetricsAction(options) {
 				type: "sdlc",
 				phase_token_breakdown: phaseBreakdownMap[run$1.run_id] ?? null
 			}));
-			const storyMetricsWithFindings = storyMetrics.map((sm) => ({
-				...sm,
-				verification_findings: findingCountsByStoryRun.get(`${sm.story_key}:${sm.run_id}`) ?? { ...ZERO_FINDING_COUNTS },
-				verification_ran: verificationRanByStoryRun.get(`${sm.story_key}:${sm.run_id}`) ?? false
-			}));
+			const storyMetricsWithFindings = storyMetrics.map((sm) => {
+				const key = `${sm.story_key}:${sm.run_id}`;
+				const counts = findingCountsByStoryRun.get(key) ?? { ...ZERO_FINDING_COUNTS };
+				const byAuthor = findingsByAuthorByStoryRun.get(key) ?? { ...ZERO_FINDINGS_BY_AUTHOR };
+				return {
+					...sm,
+					verification_findings: {
+						...counts,
+						byAuthor
+					},
+					verification_ran: verificationRanByStoryRun.get(key) ?? false,
+					probe_author: probeAuthorByStoryRun.get(key) ?? { ...ZERO_PROBE_AUTHOR_METRICS }
+				};
+			});
 			const jsonPayload = {
 				runs: runsWithBreakdown,
 				graph_runs: factoryRuns,
 				story_metrics: storyMetricsWithFindings
 			};
+			if (probeAuthorSummary) {
+				const allMetrics = storyMetricsWithFindings.map((sm) => sm.probe_author);
+				const aggregate$1 = aggregateProbeAuthorMetrics(allMetrics, storyMetricsWithFindings.length);
+				jsonPayload.probe_author_summary = aggregate$1;
+			}
 			if (doltMetrics !== void 0) if (aggregate) {
 				const aggregateResults = doltMetrics.map((m) => ({
 					task_type: m.taskType,
@@ -6100,7 +6124,7 @@ async function runMetricsAction(options) {
 	}
 }
 function registerMetricsCommand(program, _version = "0.0.0", projectRoot = process.cwd()) {
-	program.command("metrics").description("Show historical pipeline run metrics and cross-run comparison").option("--project-root <path>", "Project root directory", projectRoot).option("--output-format <format>", "Output format: human (default) or json", "human").option("--limit <n>", "Number of runs to show (default: 10)", (v) => parseInt(v, 10), 10).option("--compare <run-id-a,run-id-b>", "Compare two runs side-by-side (comma-separated IDs, e.g. abc123,def456)").option("--tag-baseline <run-id>", "Mark a run as the performance baseline").option("--analysis <run-id>", "Read and output the analysis report for the specified run (AC5 of Story 17-3)").option("--sprint <sprint>", "Filter StateStore metrics by sprint (e.g. sprint-1)").option("--story <story-key>", "Filter StateStore metrics by story key (e.g. 26-1)").option("--task-type <type>", "Filter StateStore metrics by task type (e.g. dev-story)").option("--since <iso-date>", "Filter StateStore metrics at or after this ISO timestamp").option("--aggregate", "Aggregate StateStore metrics grouped by task_type").option("--efficiency", "Show telemetry efficiency scores for recent stories").option("--recommendations", "Show all telemetry recommendations across stories").option("--turns <storyKey>", "Show per-turn analysis for a specific story").option("--consumers <storyKey>", "Show consumer stats for a specific story").option("--categories", "Show category stats (optionally scoped by --story <storyKey>)").option("--compare-stories <storyA,storyB>", "Compare efficiency scores of two stories side-by-side (comma-separated keys)").option("--routing-recommendations", "Show routing recommendations derived from phase token breakdown history").option("--run <run-id>", "Show per-iteration score history for a specific factory run").option("--factory", "Show only factory graph run metrics (excludes SDLC runs)").action(async (opts) => {
+	program.command("metrics").description("Show historical pipeline run metrics and cross-run comparison").option("--project-root <path>", "Project root directory", projectRoot).option("--output-format <format>", "Output format: human (default) or json", "human").option("--limit <n>", "Number of runs to show (default: 10)", (v) => parseInt(v, 10), 10).option("--compare <run-id-a,run-id-b>", "Compare two runs side-by-side (comma-separated IDs, e.g. abc123,def456)").option("--tag-baseline <run-id>", "Mark a run as the performance baseline").option("--analysis <run-id>", "Read and output the analysis report for the specified run (AC5 of Story 17-3)").option("--sprint <sprint>", "Filter StateStore metrics by sprint (e.g. sprint-1)").option("--story <story-key>", "Filter StateStore metrics by story key (e.g. 26-1)").option("--task-type <type>", "Filter StateStore metrics by task type (e.g. dev-story)").option("--since <iso-date>", "Filter StateStore metrics at or after this ISO timestamp").option("--aggregate", "Aggregate StateStore metrics grouped by task_type").option("--efficiency", "Show telemetry efficiency scores for recent stories").option("--recommendations", "Show all telemetry recommendations across stories").option("--turns <storyKey>", "Show per-turn analysis for a specific story").option("--consumers <storyKey>", "Show consumer stats for a specific story").option("--categories", "Show category stats (optionally scoped by --story <storyKey>)").option("--compare-stories <storyA,storyB>", "Compare efficiency scores of two stories side-by-side (comma-separated keys)").option("--routing-recommendations", "Show routing recommendations derived from phase token breakdown history").option("--run <run-id>", "Show per-iteration score history for a specific factory run").option("--factory", "Show only factory graph run metrics (excludes SDLC runs)").option("--probe-author-summary", "Print cross-run probe-author KPI aggregate (Story 60-15)").action(async (opts) => {
 		const outputFormat = opts.outputFormat === "json" ? "json" : "human";
 		let compareIds;
 		if (opts.compare !== void 0) {
@@ -6137,7 +6161,8 @@ function registerMetricsCommand(program, _version = "0.0.0", projectRoot = proce
 			...compareStoriesIds !== void 0 && { compareStories: compareStoriesIds },
 			...opts.routingRecommendations !== void 0 && { routingRecommendations: opts.routingRecommendations },
 			...opts.run !== void 0 && { run: opts.run },
-			...opts.factory !== void 0 && { factory: opts.factory }
+			...opts.factory !== void 0 && { factory: opts.factory },
+			...opts.probeAuthorSummary !== void 0 && { probeAuthorSummary: opts.probeAuthorSummary }
 		};
 		const exitCode = await runMetricsAction(metricsOpts);
 		process.exitCode = exitCode;
@@ -8440,7 +8465,7 @@ function registerProbesCommand(program) {
 			probesA = extractProbesFromArtifact(artifactA);
 		} catch (err) {
 			const msg = err instanceof Error ? err.message : String(err);
-			emitError(format, `failed to read artifact A: ${msg}`);
+			emitError$1(format, `failed to read artifact A: ${msg}`);
 			process.exitCode = 1;
 			return;
 		}
@@ -8448,7 +8473,7 @@ function registerProbesCommand(program) {
 			probesB = extractProbesFromArtifact(artifactB);
 		} catch (err) {
 			const msg = err instanceof Error ? err.message : String(err);
-			emitError(format, `failed to read artifact B: ${msg}`);
+			emitError$1(format, `failed to read artifact B: ${msg}`);
 			process.exitCode = 1;
 			return;
 		}
@@ -8463,7 +8488,7 @@ function registerProbesCommand(program) {
 		});
 	});
 }
-function emitError(format, message) {
+function emitError$1(format, message) {
 	if (format === "json") process.stdout.write(JSON.stringify({
 		success: false,
 		error: message
@@ -8633,6 +8658,83 @@ async function runProbeAuthorDispatch(opts, projectRoot, registry) {
 	return result.result === "success" ? 0 : 1;
 }
 
+//#endregion
+//#region src/cli/commands/annotate.ts
+function registerAnnotateCommand(program, _version, projectRoot) {
+	program.command("annotate").description("Post-hoc operator annotation on a verification finding (probe-author KPI confirmed-defect tagging — Story 60-15)").requiredOption("--story <key>", "Story key (e.g. 1-12)").requiredOption("--finding-category <category>", "Finding category to annotate (e.g. runtime-probe-fail, runtime-probe-error-response)").option("--probe-name <name>", "Optional probe name to narrow the annotation when multiple probes share a category").option("--note <text>", "Free-form note explaining the judgment").option("--run-id <id>", "Run ID to annotate (defaults to the latest run for this project)").option("--confirmed-defect", "Annotate as: probe failure caught a real defect").option("--false-positive", "Annotate as: probe failure was not a real defect").option("--probe-bug", "Annotate as: probe itself was buggy (not a defect catch or false positive)").option("--output-format <format>", "Output format: human (default) or json", "human").option("--project-root <path>", "Project root directory", projectRoot).action(async (opts) => {
+		const exitCode = await runAnnotateAction(opts);
+		process.exitCode = exitCode;
+	});
+}
+async function runAnnotateAction(opts) {
+	const format = opts.outputFormat === "json" ? "json" : "human";
+	const judgmentCount = (opts.confirmedDefect ? 1 : 0) + (opts.falsePositive ? 1 : 0) + (opts.probeBug ? 1 : 0);
+	if (judgmentCount !== 1) return emitError(format, `exactly one of --confirmed-defect, --false-positive, --probe-bug must be provided (got ${judgmentCount})`);
+	const judgment = opts.confirmedDefect ? "confirmed-defect" : opts.falsePositive ? "false-positive" : "probe-bug";
+	const dbRoot = await resolveMainRepoRoot(opts.projectRoot);
+	let runId = opts.runId;
+	if (runId === void 0) {
+		const adapter = createDatabaseAdapter({
+			backend: "auto",
+			basePath: opts.projectRoot
+		});
+		try {
+			const latest = await getLatestRun(adapter);
+			if (latest === null || latest === void 0) return emitError(format, "no runs found — pass --run-id explicitly");
+			runId = latest.id;
+		} finally {
+			await adapter.close();
+		}
+	}
+	const manifest = RunManifest.open(runId, join$1(dbRoot, "runs"));
+	let existingData;
+	try {
+		existingData = await manifest.read();
+	} catch (err) {
+		const msg = err instanceof Error ? err.message : String(err);
+		return emitError(format, `failed to read manifest for run ${runId}: ${msg}`);
+	}
+	const storyState = existingData.per_story_state[opts.story];
+	if (storyState === void 0) return emitError(format, `story ${opts.story} not found in run ${runId}`);
+	const verificationResult = storyState.verification_result;
+	if (verificationResult === void 0 || verificationResult === null) return emitError(format, `story ${opts.story} has no verification_result to annotate`);
+	const newAnnotation = {
+		findingCategory: opts.findingCategory,
+		judgment,
+		createdAt: new Date().toISOString(),
+		...opts.probeName !== void 0 ? { probeName: opts.probeName } : {},
+		...opts.note !== void 0 ? { note: opts.note } : {}
+	};
+	const existingAnnotations = verificationResult.annotations ?? [];
+	const updatedVerificationResult = {
+		...verificationResult,
+		annotations: [...existingAnnotations, newAnnotation]
+	};
+	try {
+		await manifest.patchStoryState(opts.story, { verification_result: updatedVerificationResult });
+	} catch (err) {
+		const msg = err instanceof Error ? err.message : String(err);
+		return emitError(format, `failed to write annotation: ${msg}`);
+	}
+	if (format === "json") process.stdout.write(JSON.stringify({
+		success: true,
+		runId,
+		storyKey: opts.story,
+		annotation: newAnnotation,
+		totalAnnotations: existingAnnotations.length + 1
+	}) + "\n");
+	else process.stdout.write(`annotated story ${opts.story} (${runId}): ${judgment} on ${opts.findingCategory}` + (opts.probeName !== void 0 ? ` [${opts.probeName}]` : "") + "\n");
+	return 0;
+}
+function emitError(format, message) {
+	if (format === "json") process.stdout.write(JSON.stringify({
+		success: false,
+		error: message
+	}) + "\n");
+	else process.stderr.write(`Error: ${message}\n`);
+	return 1;
+}
+
 //#endregion
 //#region src/cli/commands/history.ts
 function registerHistoryCommand(program) {
@@ -9238,6 +9340,7 @@ async function createProgram() {
 	registerMigrateCommand(program);
 	registerProbesCommand(program);
 	registerProbeAuthorCommand(program, version, process.cwd(), registry);
+	registerAnnotateCommand(program, version, process.cwd());
 	registerRepoMapCommand(program);
 	registerRoutingCommand(program);
 	registerCostCommand(program, version);

package/dist/{health-R9GiVIiQ.js → health-BmEu3n9Z.js}

@@ -3666,7 +3666,8 @@ const RuntimeProbeSchema = z.object({
 	timeout_ms: z.number().int().positive().optional(),
 	description: z.string().optional(),
 	expect_stdout_no_regex: z.array(z.string().min(1)).optional(),
-	expect_stdout_regex: z.array(z.string().min(1)).optional()
+	expect_stdout_regex: z.array(z.string().min(1)).optional(),
+	_authoredBy: z.enum(["probe-author", "create-story-ac-transfer"]).optional()
 });
 /** Zod schema for the full list (wrapping the per-probe schema). */
 const RuntimeProbeListSchema = z.array(RuntimeProbeSchema);
@@ -4140,8 +4141,8 @@ var RuntimeProbeCheck = class {
 		if (context.sourceEpicContent !== void 0) {
 			if (detectsEventDrivenAC(context.sourceEpicContent) && !probesInvokeProductionTrigger(parsed.probes)) findings.push({
 				category: CATEGORY_MISSING_TRIGGER,
-				severity: "warn",
-				message: "source AC describes an event-driven mechanism (hook / timer / signal / webhook) but no probe's command invokes a known production trigger (git merge/pull/push, systemctl, crontab, kill -<sig>, curl -X POST, etc.). Probes that call the implementation directly skip the wiring layer the AC's user-facing event would exercise — see strata Run 13 / Story 1-12 for the canonical case (post-merge hook never fires under git's conflict semantic). Authoring guidance: probes/event-driven section of create-story.md."
+				severity: "error",
+				message: "source AC describes an event-driven mechanism (hook / timer / signal / webhook) but no probe's command invokes a known production trigger (git merge/pull/push, systemctl, crontab, kill -<sig>, curl -X POST, etc.). Probes that call the implementation directly skip the wiring layer the AC's user-facing event would exercise — see strata Run 13 / Story 1-12 for the canonical case (post-merge hook never fires under git's conflict semantic). Authoring guidance: probes/event-driven section of create-story.md, or invoke probe-author to derive AC-grounded probes automatically (Epic 60 Phase 2)."
 			});
 		}
 		for (const probe of parsed.probes) {
@@ -4170,7 +4171,8 @@ var RuntimeProbeCheck = class {
 				...result.exitCode !== void 0 ? { exitCode: result.exitCode } : {},
 				stdoutTail: result.stdoutTail,
 				stderrTail: result.stderrTail,
-				durationMs: result.durationMs
+				durationMs: result.durationMs,
+				_authoredBy: probe._authoredBy ?? "create-story-ac-transfer"
 			});
 		}
 		const status = findings.some((f) => f.severity === "error") ? "fail" : findings.some((f) => f.severity === "warn") ? "warn" : "pass";
@@ -4508,10 +4510,90 @@ function computeTakenOptionPerGroup(hardClauses, workingDir) {
 *   3. The presence of `## Runtime Probes` heading followed by a fenced yaml block
 *      (represented as a single "runtime-probes-section" clause)
 */
+/**
+* Negation phrases that mark a paragraph as "the paths in this paragraph
+* are references the dev should NOT deliver/modify". When any of these
+* phrases appears in a paragraph, every path-clause in that paragraph is
+* tagged `negation: true` so the verification emit routes them to the
+* info-severity `source-ac-negation-reference` finding instead of the
+* under-delivery error path.
+*
+* Strata Run 16 (Story 1-16, 2026-04-27): the AC contained
+*
+*   "the existing test scaffolding is documented (NOT replaced):
+*    `packages/memory` already uses vitest (Story 1.8+);
+*    `packages/memory-mcp` already uses pytest (Story 1.10+); ...
+*    1.16 does NOT replace or rewrite existing test infrastructure."
+*
+* Substrate emitted 6 ERROR-level under-delivery findings on the listed
+* paths; the dev had correctly NOT modified them. The flood of
+* false-positive ERRORs masked a real WARN about the missing `## Runtime
+* Probes` section (which itself would have surfaced two real defects in
+* the delivery). Detecting the negation context up-front separates the
+* legitimate "paths the AC mentions but the dev should not modify" case
+* from genuine under-delivery.
+*
+* Patterns are case-sensitive on the keyword (NOT, MUST NOT, gitignored)
+* because lowercased forms are common in non-imperative prose ("not
+* really" / "must not exceed" appear in unrelated contexts and
+* over-trigger). The strata observation enumerates the canonical forms.
+*/
+const NEGATION_PHRASE_PATTERNS = [
+	/\(NOT\s+\w+/,
+	/documented\s*\(NOT/i,
+	/\bMUST\s+NOT\b|\bSHALL\s+NOT\b/,
+	/\bdo(?:es)?\s+NOT\s+\w+/,
+	/\bdeferred\s+to\b/i,
+	/\b(?:is|are)\s+gitignored\b/i
+];
+/**
+* Find all line indices that fall within a negation context. The scope of
+* a single negation context is the line that contains the negation phrase
+* PLUS any markdown indented-continuation lines following it (so a bullet
+* that wraps onto multiple indented lines is treated as one logical unit).
+*
+* The continuation walk stops at the first of: blank line, next markdown
+* bullet (`- `, `* `, numbered list), or any non-indented non-blank line.
+*
+* Coarser scopes (e.g., paragraph-wide aggregation) over-triggered on
+* test fixtures like:
+*
+*   The implementation MUST validate input.
+*   The system MUST NOT skip authentication.
+*   Files SHALL be placed in `src/auth/validator.ts`.
+*
+* — three independent statements wrapped without blank-line separators.
+* The "MUST NOT" on line 2 should not mark the unrelated path on line 3
+* as a negation reference. Only the bullet's CONTINUATION-LINE structure
+* constitutes "the same logical reference unit" as the negation phrase.
+*/
+function detectNegationContextLines(lines) {
+	const result = new Set();
+	for (let i = 0; i < lines.length; i++) {
+		const line = lines[i] ?? "";
+		if (NEGATION_PHRASE_PATTERNS.some((pat) => pat.test(line))) {
+			result.add(i);
+			let j = i + 1;
+			while (j < lines.length) {
+				const next = lines[j] ?? "";
+				if (next.trim() === "") break;
+				if (/^\s*(?:-|\*|\d+\.)\s+/.test(next)) break;
+				if (/^\s+\S/.test(next)) {
+					result.add(j);
+					j++;
+					continue;
+				}
+				break;
+			}
+		}
+	}
+	return result;
+}
 function extractHardClauses(sectionContent) {
 	const clauses = [];
 	const lines = sectionContent.split("\n");
 	const alternativeOptions = detectAlternativeOptions(lines);
+	const negationContextLines = detectNegationContextLines(lines);
 	const mustPattern = /\b(MUST NOT|MUST|SHALL NOT|SHALL)\b/;
 	for (const line of lines) {
 		const match = mustPattern.exec(line);
@@ -4530,10 +4612,12 @@ function extractHardClauses(sectionContent) {
 		let pathMatch;
 		while ((pathMatch = pathPattern.exec(line)) !== null) {
 			const alt = findOptionForLine(lineIdx, alternativeOptions);
+			const inNegation = negationContextLines.has(lineIdx);
 			clauses.push({
 				type: "path",
 				text: `\`${pathMatch[1]}\``,
-				...alt ? { alternative: alt } : {}
+				...alt ? { alternative: alt } : {},
+				...inNegation ? { negation: true } : {}
 			});
 		}
 	}
@@ -4583,15 +4667,24 @@ var SourceAcFidelityCheck = class {
 		for (const clause of hardClauses) if (clause.type === "runtime-probes-section") {
 			if (!storyContent.includes("## Runtime Probes")) {
 				const truncated = clause.text.length > 120 ? clause.text.slice(0, 120) : clause.text;
+				const isEventDrivenAc = detectsEventDrivenAC(context.sourceEpicContent);
 				findings.push({
 					category: "source-ac-drift",
-					severity: "warn",
-					message: `runtime-probes-section: "${truncated}" present in epics source but absent in story artifact`
+					severity: isEventDrivenAc ? "error" : "warn",
+					message: isEventDrivenAc ? `runtime-probes-section: "${truncated}" present in epics source but absent in story artifact AND source AC is event-driven (probes are required for event-driven ACs — runtime-probes check will skip without the section)` : `runtime-probes-section: "${truncated}" present in epics source but absent in story artifact`
 				});
 			}
 		} else if (!storyContent.includes(clause.text)) {
 			const truncated = clause.text.length > 120 ? clause.text.slice(0, 120) : clause.text;
 			if (clause.type === "path") {
+				if (clause.negation === true) {
+					findings.push({
+						category: "source-ac-negation-reference",
+						severity: "info",
+						message: `path: "${truncated}" referenced in source AC inside a negation context (e.g., "(NOT replaced)", "MUST NOT", "deferred to", "documented (NOT", "does NOT replace", "is gitignored") — the AC explicitly directed the dev NOT to deliver/modify this path; treated as reference-only, not a deliverable`
+					});
+					continue;
+				}
 				if (isOperationalPath(clause.text)) {
 					findings.push({
 						category: "source-ac-operational-path-reference",
@@ -4821,7 +4914,8 @@ const StoredVerificationFindingSchema = z.object({
 	exitCode: z.number().int().optional(),
 	stdoutTail: z.string().optional(),
 	stderrTail: z.string().optional(),
-	durationMs: z.number().nonnegative().optional()
+	durationMs: z.number().nonnegative().optional(),
+	_authoredBy: z.enum(["probe-author", "create-story-ac-transfer"]).optional()
 });
 /**
 * Schema for a single per-check verification result stored in the manifest.
@@ -4846,6 +4940,30 @@ const StoredVerificationCheckResultSchema = z.object({
 * Mirrors VerificationSummary from packages/sdlc/src/verification/types.ts
 * without importing from that module (avoids circular dependency).
 */
+/**
+* Story 60-15: an operator-added annotation on a verification finding,
+* recording the post-hoc judgment about whether a probe failure was
+* a real defect catch, a false positive, or a probe bug.
+*
+* Persisted under `per_story_state[storyKey].verification_result.annotations[]`.
+* Set via `substrate annotate --story <key> --finding-category <cat>
+* --confirmed-defect|--false-positive|--probe-bug`.
+*
+* Used by `rollupProbeAuthorMetrics` to compute the catch-rate KPI's
+* `confirmed-defect` count: a probe-author probe failure that the
+* operator subsequently confirmed as catching a real defect.
+*/
+const StoredVerificationAnnotationSchema = z.object({
+	findingCategory: z.string(),
+	judgment: z.enum([
+		"confirmed-defect",
+		"false-positive",
+		"probe-bug"
+	]),
+	probeName: z.string().optional(),
+	note: z.string().optional(),
+	createdAt: z.string()
+});
 const StoredVerificationSummarySchema = z.object({
 	storyKey: z.string(),
 	checks: z.array(StoredVerificationCheckResultSchema),
@@ -4854,7 +4972,8 @@ const StoredVerificationSummarySchema = z.object({
 		"warn",
 		"fail"
 	]),
-	duration_ms: z.number().nonnegative()
+	duration_ms: z.number().nonnegative(),
+	annotations: z.array(StoredVerificationAnnotationSchema).optional()
 });
 
 //#endregion
@@ -5599,6 +5718,171 @@ function rollupFindingCounts(summary) {
 	};
 }
 
+//#endregion
+//#region packages/sdlc/dist/run-model/probe-author-metrics.js
+/**
+* Probe-author per-story rollup — Story 60-15.
+*
+* Collapses every probe-author-relevant signal in a
+* StoredVerificationSummary (including any operator annotations on
+* probe-failures) into a flat `{dispatched, probesAuthoredCount,
+* authoredProbesFailedCount, authoredProbesCaughtConfirmedDefectCount}`
+* shape suitable for per-story surfacing in `substrate status`/`metrics`
+* JSON output and for cross-run aggregation in the `--probe-author-summary`
+* flag.
+*
+* Intentionally pure (mirrors `rollupFindingCounts` shape): no I/O, no
+* logger, no throw. Backward-compat is load-bearing — pre-60-15 manifests
+* have no `_authoredBy` discriminator on their stored findings, no
+* `annotations` array, and probe-author wasn't actually running (Sprint
+* 13/Sprint 20 lesson — the manifest registration bug). The rollup must
+* produce sensible zero values on every legacy code path.
+*/
+/** Zero rollup, used as default when no probe-author signal is present. */
+const ZERO_PROBE_AUTHOR_METRICS = Object.freeze({
+	dispatched: false,
+	probesAuthoredCount: 0,
+	authoredProbesFailedCount: 0,
+	authoredProbesCaughtConfirmedDefectCount: 0
+});
+const ZERO_FINDINGS_BY_AUTHOR = Object.freeze({
+	"probe-author": {
+		error: 0,
+		warn: 0,
+		info: 0
+	},
+	"create-story-ac-transfer": {
+		error: 0,
+		warn: 0,
+		info: 0
+	}
+});
+/**
+* Compute the per-story probe-author rollup from a verification summary.
+*
+* Inputs:
+*  - `summary`: the per-story stored verification summary (`per_story_state
+*    [storyKey].verification_result` in run manifest terms)
+*  - `dispatchedHint`: optional override — when the caller has direct
+*    knowledge that probe-author dispatched (e.g., from a captured
+*    `probe-author:appended-to-artifact` event on a fresh run), pass true.
+*    Useful when probe-author authored probes that all passed (no
+*    failure findings to attribute, but presence is real).
+*
+* Backward-compat: undefined/null summary → zero rollup. Missing
+* `_authoredBy` field on findings → counted under
+* `'create-story-ac-transfer'` per the schema's documented semantic.
+* Missing `annotations` array → confirmed-defect count is 0.
+*/
+function rollupProbeAuthorMetrics(summary, dispatchedHint) {
+	if (summary === void 0 || summary === null) return dispatchedHint === true ? {
+		...ZERO_PROBE_AUTHOR_METRICS,
+		dispatched: true
+	} : { ...ZERO_PROBE_AUTHOR_METRICS };
+	const authoredFailures = [];
+	for (const check of summary.checks) {
+		const findings = check.findings ?? [];
+		for (const f of findings) {
+			const author = f._authoredBy ?? "create-story-ac-transfer";
+			if (author !== "probe-author") continue;
+			if (!f.category.startsWith("runtime-probe-")) continue;
+			const probeName = extractProbeName(f.message) ?? f.category;
+			authoredFailures.push({
+				probeName,
+				category: f.category
+			});
+		}
+	}
+	const annotations = summary.annotations ?? [];
+	let confirmedDefectCount = 0;
+	for (const failure of authoredFailures) {
+		const matched = annotations.find((a) => a.judgment === "confirmed-defect" && a.findingCategory === failure.category && (a.probeName === void 0 || a.probeName === failure.probeName));
+		if (matched !== void 0) confirmedDefectCount += 1;
+	}
+	return {
+		dispatched: dispatchedHint === true || authoredFailures.length > 0,
+		probesAuthoredCount: authoredFailures.length,
+		authoredProbesFailedCount: authoredFailures.length,
+		authoredProbesCaughtConfirmedDefectCount: confirmedDefectCount
+	};
+}
+/**
+* Compute the byAuthor breakdown of finding counts on a verification summary.
+* Each finding's `_authoredBy` (default `'create-story-ac-transfer'` when
+* absent) routes its severity into the appropriate per-author bucket.
+*/
+function rollupFindingsByAuthor(summary) {
+	const result = {
+		"probe-author": {
+			error: 0,
+			warn: 0,
+			info: 0
+		},
+		"create-story-ac-transfer": {
+			error: 0,
+			warn: 0,
+			info: 0
+		}
+	};
+	if (summary === void 0 || summary === null) return result;
+	for (const check of summary.checks) {
+		const findings = check.findings ?? [];
+		for (const f of findings) {
+			const author = f._authoredBy ?? "create-story-ac-transfer";
+			const bucket = result[author];
+			switch (f.severity) {
+				case "error":
+					bucket.error += 1;
+					break;
+				case "warn":
+					bucket.warn += 1;
+					break;
+				case "info":
+					bucket.info += 1;
+					break;
+			}
+		}
+	}
+	return result;
+}
+/**
+* Sum per-story probe-author rollups into a single aggregate. `totalStories`
+* is the denominator for `probeAuthorDispatchedPct` (stories where dispatch
+* could have happened, regardless of outcome). Pass the count of stories
+* the run touched, NOT just the ones with successful dispatch.
+*/
+function aggregateProbeAuthorMetrics(perStory, totalStories) {
+	let dispatched = 0;
+	let authored = 0;
+	let failed = 0;
+	let confirmed = 0;
+	for (const story of perStory) {
+		if (story.dispatched) dispatched += 1;
+		authored += story.probesAuthoredCount;
+		failed += story.authoredProbesFailedCount;
+		confirmed += story.authoredProbesCaughtConfirmedDefectCount;
+	}
+	return {
+		totalStoriesDispatched: totalStories,
+		probeAuthorDispatchedCount: dispatched,
+		probeAuthorDispatchedPct: totalStories > 0 ? dispatched / totalStories : 0,
+		totalAuthoredProbes: authored,
+		totalAuthoredProbesFailed: failed,
+		totalConfirmedDefectsCaught: confirmed,
+		catchRateByCount: authored > 0 ? failed / authored : 0,
+		catchRateByConfirmedDefect: authored > 0 ? confirmed / authored : 0
+	};
+}
+/** Extract the probe name from a runtime-probe finding's message. The
+*  runtime-probe-check formatter writes `probe "<name>"...` as the leading
+*  pattern across all category branches (fail/timeout/assertion-fail/
+*  error-response). Returns undefined when the message doesn't match
+*  (fault tolerance — the rollup falls back to category-as-name). */
+function extractProbeName(message) {
+	const match = /^probe\s+"([^"]+)"/.exec(message);
+	return match?.[1];
+}
+
 //#endregion
 //#region packages/sdlc/dist/run-model/supervisor-lock.js
 const defaultLogger = console;
@@ -6343,5 +6627,5 @@ function registerHealthCommand(program, _version = "0.0.0", projectRoot = proces
 }
 
 //#endregion
-export { BMAD_BASELINE_TOKENS_FULL, DEFAULT_STALL_THRESHOLD_SECONDS, DoltMergeConflict, FileStateStore, FindingsInjector, RunManifest, RuntimeProbeListSchema, STOP_AFTER_VALID_PHASES, STORY_KEY_PATTERN$1 as STORY_KEY_PATTERN, SUBSTRATE_OWNED_SETTINGS_KEYS, SupervisorLock, VALID_PHASES, WorkGraphRepository, ZERO_FINDING_COUNTS, __commonJS, __require, __toESM, applyConfigToGraph, buildPipelineStatusOutput, createDatabaseAdapter$1 as createDatabaseAdapter, createDefaultVerificationPipeline, createGraphOrchestrator, createSdlcCodeReviewHandler, createSdlcCreateStoryHandler, createSdlcDevStoryHandler, createSdlcPhaseHandler, createStateStore, detectCycles, detectsEventDrivenAC, extractTargetFilesFromStoryContent, findPackageRoot, formatOutput, formatPipelineStatusHuman, formatPipelineSummary, formatTokenTelemetry, getAllDescendantPids, getAutoHealthData, getSubstrateDefaultSettings, inspectProcessTree, isOrchestratorProcessLine, parseDbTimestampAsUtc, parseRuntimeProbes, registerHealthCommand, renderFindings, resolveBmadMethodSrcPath, resolveBmadMethodVersion, resolveGraphPath, resolveMainRepoRoot, resolveRunManifest, rollupFindingCounts, runHealthAction, validateStoryKey };
-//# sourceMappingURL=health-R9GiVIiQ.js.map
+export { BMAD_BASELINE_TOKENS_FULL, DEFAULT_STALL_THRESHOLD_SECONDS, DoltMergeConflict, FileStateStore, FindingsInjector, RunManifest, RuntimeProbeListSchema, STOP_AFTER_VALID_PHASES, STORY_KEY_PATTERN$1 as STORY_KEY_PATTERN, SUBSTRATE_OWNED_SETTINGS_KEYS, SupervisorLock, VALID_PHASES, WorkGraphRepository, ZERO_FINDINGS_BY_AUTHOR, ZERO_FINDING_COUNTS, ZERO_PROBE_AUTHOR_METRICS, __commonJS, __require, __toESM, aggregateProbeAuthorMetrics, applyConfigToGraph, buildPipelineStatusOutput, createDatabaseAdapter$1 as createDatabaseAdapter, createDefaultVerificationPipeline, createGraphOrchestrator, createSdlcCodeReviewHandler, createSdlcCreateStoryHandler, createSdlcDevStoryHandler, createSdlcPhaseHandler, createStateStore, detectCycles, detectsEventDrivenAC, extractTargetFilesFromStoryContent, findPackageRoot, formatOutput, formatPipelineStatusHuman, formatPipelineSummary, formatTokenTelemetry, getAllDescendantPids, getAutoHealthData, getSubstrateDefaultSettings, inspectProcessTree, isOrchestratorProcessLine, parseDbTimestampAsUtc, parseRuntimeProbes, registerHealthCommand, renderFindings, resolveBmadMethodSrcPath, resolveBmadMethodVersion, resolveGraphPath, resolveMainRepoRoot, resolveRunManifest, rollupFindingCounts, rollupFindingsByAuthor, rollupProbeAuthorMetrics, runHealthAction, validateStoryKey };
+//# sourceMappingURL=health-BmEu3n9Z.js.map

package/dist/{health-C-ZrHJwL.js → health-CAV-PNE4.js}

@@ -1,4 +1,4 @@
-import { DEFAULT_STALL_THRESHOLD_SECONDS, getAllDescendantPids, getAutoHealthData, inspectProcessTree, isOrchestratorProcessLine, registerHealthCommand, runHealthAction } from "./health-R9GiVIiQ.js";
+import { DEFAULT_STALL_THRESHOLD_SECONDS, getAllDescendantPids, getAutoHealthData, inspectProcessTree, isOrchestratorProcessLine, registerHealthCommand, runHealthAction } from "./health-BmEu3n9Z.js";
 import "./logger-KeHncl-f.js";
 import "./dist-VcMmfo2w.js";
 import "./decisions-C0pz9Clx.js";

package/dist/index.d.ts

@@ -2117,6 +2117,58 @@ interface OrchestratorEvents {
     mode: 'enabled' | 'disabled';
     source: 'cli' | 'env' | 'default';
   };
+  /**
+   * Story 60-15: probe-author dispatch lifecycle events. Each event fires
+   * exactly once per probe-author phase invocation (or zero times when the
+   * phase didn't run). Powers the per-story telemetry breakdown surfaced
+   * via `substrate status`/`metrics` and the cross-run probe-author
+   * catch-rate KPI.
+   */
+  /** Probe-author dispatch completed for a story (success or skip-after-
+   *  re-read). Existing event emitted by 60-13's runProbeAuthor; 60-15
+   *  formalizes the schema as part of the lifecycle event family. */
+  'probe-author:dispatched': {
+    storyKey: string;
+    runId: string;
+    probesAuthoredCount: number;
+    dispatchDurationMs: number;
+    costUsd: number;
+  };
+  /** Probe-author agent's YAML output successfully parsed. Counts probes
+   *  authored before any append/idempotency check. */
+  'probe-author:output-parsed': {
+    storyKey: string;
+    runId: string;
+    probesParsedCount: number;
+  };
+  /** Probe-author probes appended to the story artifact. The terminal
+   *  success event for the phase. `probesAuthoredCount` is the count of
+   *  NEW probes added (excluding any that were already present). */
+  'probe-author:appended-to-artifact': {
+    storyKey: string;
+    runId: string;
+    probesAuthoredCount: number;
+    storyFilePath: string;
+  };
+  /** Probe-author phase skipped (gate fired). Distinguishes the two skip
+   *  reasons so operators can debug the gating decision. */
+  'probe-author:skipped': {
+    storyKey: string;
+    runId: string;
+    reason: 'non-event-driven' | 'author-declared-probes-present';
+  };
+  /** A probe carrying `_authoredBy: 'probe-author'` failed at runtime. Emitted
+   *  by runtime-probe-check when it produces a probe-failure finding whose
+   *  source probe was authored by probe-author. NOTE: probe-failure is NOT
+   *  the same as defect-caught — operators tag confirmed-defect via
+   *  `substrate annotate`. This event surfaces failures regardless of
+   *  annotation status; the rollup helper distinguishes them downstream. */
+  'probe-author:authored-probe-failed': {
+    storyKey: string;
+    runId: string;
+    probeName: string;
+    findingCategory: string;
+  };
   /**
    * Story 62-3: code-review agent emitted YAML output that failed schema
    * validation (typically a parse error from unquoted-colon-in-value or

package/dist/{run-CL1C4M2f.js → run-CXTAvjCK.js}

@@ -1,8 +1,8 @@
-import "./health-R9GiVIiQ.js";
+import "./health-BmEu3n9Z.js";
 import "./logger-KeHncl-f.js";
 import "./helpers-CElYrONe.js";
 import "./dist-VcMmfo2w.js";
-import { normalizeGraphSummaryToStatus, registerRunCommand, resolveMaxReviewCycles, runRunAction, wireNdjsonEmitter } from "./run-Dm70k8bR.js";
+import { normalizeGraphSummaryToStatus, registerRunCommand, resolveMaxReviewCycles, runRunAction, wireNdjsonEmitter } from "./run-DnwsCYYA.js";
 import "./routing-CcBOCuC9.js";
 import "./decisions-C0pz9Clx.js";
 

package/dist/{run-Dm70k8bR.js → run-DnwsCYYA.js}

@@ -1,4 +1,4 @@
-import { BMAD_BASELINE_TOKENS_FULL, DoltMergeConflict, FileStateStore, FindingsInjector, RunManifest, RuntimeProbeListSchema, STOP_AFTER_VALID_PHASES, STORY_KEY_PATTERN, VALID_PHASES, WorkGraphRepository, __commonJS, __require, __toESM, applyConfigToGraph, buildPipelineStatusOutput, createDatabaseAdapter, createDefaultVerificationPipeline, createGraphOrchestrator, createSdlcCodeReviewHandler, createSdlcCreateStoryHandler, createSdlcDevStoryHandler, createSdlcPhaseHandler, detectCycles, detectsEventDrivenAC, extractTargetFilesFromStoryContent, formatOutput, formatPipelineSummary, formatTokenTelemetry, inspectProcessTree, parseDbTimestampAsUtc, renderFindings, resolveGraphPath, resolveMainRepoRoot, validateStoryKey } from "./health-R9GiVIiQ.js";
+import { BMAD_BASELINE_TOKENS_FULL, DoltMergeConflict, FileStateStore, FindingsInjector, RunManifest, RuntimeProbeListSchema, STOP_AFTER_VALID_PHASES, STORY_KEY_PATTERN, VALID_PHASES, WorkGraphRepository, __commonJS, __require, __toESM, applyConfigToGraph, buildPipelineStatusOutput, createDatabaseAdapter, createDefaultVerificationPipeline, createGraphOrchestrator, createSdlcCodeReviewHandler, createSdlcCreateStoryHandler, createSdlcDevStoryHandler, createSdlcPhaseHandler, detectCycles, detectsEventDrivenAC, extractTargetFilesFromStoryContent, formatOutput, formatPipelineSummary, formatTokenTelemetry, inspectProcessTree, parseDbTimestampAsUtc, renderFindings, resolveGraphPath, resolveMainRepoRoot, validateStoryKey } from "./health-BmEu3n9Z.js";
 import { createLogger } from "./logger-KeHncl-f.js";
 import { TypedEventBusImpl, createEventBus, createTuiApp, isTuiCapable, printNonTtyWarning, sleep } from "./helpers-CElYrONe.js";
 import { ADVISORY_NOTES, Categorizer, ConsumerAnalyzer, DEFAULT_GLOBAL_SETTINGS, DispatcherImpl, DoltClient, ESCALATION_DIAGNOSIS, EXPERIMENT_RESULT, EfficiencyScorer, IngestionServer, LogTurnAnalyzer, OPERATIONAL_FINDING, Recommender, RoutingRecommender, RoutingResolver, RoutingTelemetry, RoutingTokenAccumulator, RoutingTuner, STORY_METRICS, STORY_OUTCOME, SubstrateConfigSchema, TEST_EXPANSION_FINDING, TEST_PLAN, TelemetryNormalizer, TelemetryPipeline, TurnAnalyzer, addTokenUsage, aggregateTokenUsageForRun, aggregateTokenUsageForStory, callLLM, createConfigSystem, createDatabaseAdapter$1, createDecision, createPipelineRun, createRequirement, detectInterfaceChanges, getArtifactByTypeForRun, getArtifactsByRun, getDecisionsByCategory, getDecisionsByPhase, getDecisionsByPhaseForRun, getLatestRun, getPipelineRunById, getRunMetrics, getRunningPipelineRuns, getStoryMetricsForRun, getTokenUsageSummary, initSchema, listRequirements, loadModelRoutingConfig, registerArtifact, updatePipelineRun, updatePipelineRunConfig, upsertDecision, writeRunMetrics, writeStoryMetrics } from "./dist-VcMmfo2w.js";
@@ -8353,6 +8353,11 @@ async function runProbeAuthor(deps, params) {
 	};
 	if (bypassGates !== true && !detectsEventDrivenAC(epicContent)) {
 		logger$14.debug({ storyKey }, "probe-author: source AC not event-driven — skipping");
+		emitEvent?.("probe-author:skipped", {
+			storyKey,
+			runId: pipelineRunId,
+			reason: "non-event-driven"
+		});
 		return makeSkippedResult(tokenUsage, start);
 	}
 	let storyContent;
@@ -8360,6 +8365,11 @@ async function runProbeAuthor(deps, params) {
 		storyContent = await readFile$1(storyFilePath, "utf-8");
 		if (bypassGates !== true && /^## Runtime Probes/m.test(storyContent)) {
 			logger$14.info({ storyKey }, "probe-author: story artifact already has ## Runtime Probes — skipping");
+			emitEvent?.("probe-author:skipped", {
+				storyKey,
+				runId: pipelineRunId,
+				reason: "author-declared-probes-present"
+			});
 			return makeSkippedResult(tokenUsage, start);
 		}
 	} catch (err) {
@@ -8516,7 +8526,15 @@ async function runProbeAuthor(deps, params) {
 		}, "probe-author: probes failed RuntimeProbeListSchema — falling through");
 		return makeFailedResult(`schema_validation_failed: ${validationError}`, tokenUsage, start);
 	}
-	const probes = probeValidation.data;
+	const probes = probeValidation.data.map((p) => ({
+		...p,
+		_authoredBy: "probe-author"
+	}));
+	emitEvent?.("probe-author:output-parsed", {
+		storyKey,
+		runId: pipelineRunId,
+		probesParsedCount: probes.length
+	});
 	if (probes.length === 0) {
 		logger$14.info({ storyKey }, "probe-author: authored empty probes list — no probes needed");
 		emitEvent?.("probe-author:no-probes-authored", {
@@ -8555,6 +8573,12 @@ async function runProbeAuthor(deps, params) {
 			storyKey,
 			probesCount: probes.length
 		}, "probe-author: appended ## Runtime Probes section");
+		emitEvent?.("probe-author:appended-to-artifact", {
+			storyKey,
+			runId: pipelineRunId,
+			probesAuthoredCount: probes.length,
+			storyFilePath
+		});
 	} catch (err) {
 		const error = err instanceof Error ? err.message : String(err);
 		logger$14.warn({
@@ -45151,4 +45175,4 @@ function registerRunCommand(program, _version = "0.0.0", projectRoot = process.c
 
 //#endregion
 export { AdapterTelemetryPersistence, AppError, DoltRepoMapMetaRepository, DoltSymbolRepository, ERR_REPO_MAP_STORAGE_WRITE, EpicIngester, GitClient, GrammarLoader, RepoMapInjector, RepoMapModule, RepoMapQueryEngine, RepoMapStorage, SymbolParser, createContextCompiler, createDispatcher, createEventEmitter, createImplementationOrchestrator, createPackLoader, createPhaseOrchestrator, createStopAfterGate, createTelemetryAdvisor, formatPhaseCompletionSummary, getFactoryRunSummaries, getScenarioResultsForRun, getTwinRunsForRun, listGraphRuns, normalizeGraphSummaryToStatus, registerExportCommand, registerFactoryCommand, registerRunCommand, registerScenariosCommand, resolveMaxReviewCycles, resolveStoryKeys, runAnalysisPhase, runPlanningPhase, runProbeAuthor, runRunAction, runSolutioningPhase, validateStopAfterFromConflict, wireNdjsonEmitter };
-//# sourceMappingURL=run-Dm70k8bR.js.map
+//# sourceMappingURL=run-DnwsCYYA.js.map

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "substrate-ai",
-  "version": "0.20.39",
+  "version": "0.20.41",
   "description": "Substrate — multi-agent orchestration daemon for AI coding agents",
   "type": "module",
   "license": "MIT",

package/packs/bmad/prompts/create-story.md

@@ -206,6 +206,13 @@ Strata Run 13 (Story 1-12, post-merge git hook) shipped SHIP_IT after the dev's
 
 Note this example, taken to production, would have caught the strata 1-12 bug at runtime-probe phase rather than only at e2e smoke pass. That's the standard 60-10 sets.
 
+**This is now a blocking gate (Story 60-16, v0.20.41+).** Probes for stories whose AC describes an event-driven mechanism (hook, timer, signal, webhook) MUST invoke the production trigger via a known command pattern (`git merge|pull|push`, `systemctl start <unit>`, `crontab`, `kill -<signal>`, `curl -X POST`, etc.). When the gate fires (`runtime-probe-missing-production-trigger` finding category, error severity), verification status becomes `fail` and the story cannot SHIP_IT until probes invoke the trigger. Two ways to satisfy:
+
+1. **Author trigger-invoking probes directly** in this `## Runtime Probes` section — exercise the production trigger as the example above does.
+2. **Let probe-author derive probes** from the AC (Epic 60 Phase 2). When probe-author dispatches for an event-driven AC, it authors AC-grounded probes that exercise production triggers by design. The eval-validated catch rate on the v1 corpus was 4/4 (100%) under v0.20.39+. The probes carry `_authoredBy: 'probe-author'` metadata for telemetry attribution.
+
+Pre-Sprint-22 (warn-severity advisory) the gate produced false negatives at SHIP_IT time. Post-flip, the gate is the load-bearing line of defense for the trigger-invocation property.
+
 ### Examples by artifact class
 
 **Systemd unit:**