@project-ava-supernova/core 0.2.71

package/dist/dataset/capture.js

@@ -0,0 +1,203 @@
+/**
+ * Dataset Capture — silently records Ava interactions as training data.
+ *
+ * Writes JSONL to ~/.ava/datasets/raw/ — one line per interaction.
+ * Config read from ~/.ava/datasets/config.json (written by the Hub).
+ * All local. Nothing leaves the machine.
+ */
+import { join } from 'node:path';
+import { mkdir, readFile, appendFile } from 'node:fs/promises';
+import { randomUUID } from 'node:crypto';
+import { AVA_HOME } from '../core/constants.js';
+import { getTextContent } from '../core/types.js';
+import { logger } from '../core/logger.js';
+// ─── Paths ──────────────────────────────────────────────────────────────────
+const DATASET_DIR = join(AVA_HOME, 'datasets');
+const RAW_DIR = join(DATASET_DIR, 'raw');
+const CONFIG_PATH = join(DATASET_DIR, 'config.json');
+// ─── Config ─────────────────────────────────────────────────────────────────
+const DEFAULT_CONFIG = {
+    enabled: false,
+    capture_modes: ['work', 'plan', 'chat', 'teach', 'security', 'brainstorm'],
+    auto_label_corrections: true,
+    min_prompt_length: 10,
+};
+let configCache = null;
+let configLastRead = 0;
+const CONFIG_TTL = 30_000; // re-read config every 30s max
+async function loadConfig() {
+    const now = Date.now();
+    if (configCache && now - configLastRead < CONFIG_TTL)
+        return configCache;
+    try {
+        const raw = await readFile(CONFIG_PATH, 'utf-8');
+        configCache = { ...DEFAULT_CONFIG, ...JSON.parse(raw) };
+        configLastRead = now;
+        return configCache;
+    }
+    catch {
+        // No config file yet — collection disabled by default
+        configCache = DEFAULT_CONFIG;
+        configLastRead = now;
+        return DEFAULT_CONFIG;
+    }
+}
+// ─── Mode Detection ─────────────────────────────────────────────────────────
+const MODE_PREFIXES = [
+    ['[Chat Mode]', 'chat'],
+    ['[Teach Mode]', 'teach'],
+    ['[Security Mode]', 'security'],
+    ['[Brainstorm Mode]', 'brainstorm'],
+    ['[Plan Mode]', 'plan'],
+];
+function detectMode(messages) {
+    // Check the last user message (before internal planning messages) for mode prefix
+    for (let i = messages.length - 1; i >= 0; i--) {
+        const msg = messages[i];
+        if (msg.role !== 'user')
+            continue;
+        // Skip internal planning injections
+        const text = getTextContent(msg.content);
+        if (text.startsWith('[Internal Planning'))
+            continue;
+        for (const [prefix, mode] of MODE_PREFIXES) {
+            if (text.startsWith(prefix))
+                return mode;
+        }
+        // No prefix = work mode (default)
+        return 'work';
+    }
+    return 'work';
+}
+// ─── Tool Extraction ────────────────────────────────────────────────────────
+function extractToolsUsed(messages) {
+    const tools = new Set();
+    for (const msg of messages) {
+        if (msg.role === 'assistant' && 'tool_calls' in msg && msg.tool_calls) {
+            for (const tc of msg.tool_calls) {
+                tools.add(tc.function.name);
+            }
+        }
+    }
+    return Array.from(tools);
+}
+// ─── Verification Detection ─────────────────────────────────────────────────
+/** Did Ava use a read/search tool before answering? (sign of verification) */
+function didVerifyBeforeAnswering(messages) {
+    const verificationTools = new Set([
+        'file_read', 'grep', 'glob', 'list_directory', 'find_symbol',
+        'web_search', 'http_request', 'browser', 'project_index',
+        'git_status', 'git_diff',
+    ]);
+    const tools = extractToolsUsed(messages);
+    return tools.some(t => verificationTools.has(t));
+}
+// ─── Correction Detection ───────────────────────────────────────────────────
+/** Check if the latest user message looks like a correction of the previous response. */
+function detectCorrection(messages) {
+    // Need at least: user → assistant → user (correction)
+    if (messages.length < 3)
+        return null;
+    const lastUser = messages[messages.length - 1];
+    const prevAssistant = messages[messages.length - 2];
+    const prevPrevUser = messages[messages.length - 3];
+    if (lastUser?.role !== 'user' || prevAssistant?.role !== 'assistant' || prevPrevUser?.role !== 'user') {
+        return null;
+    }
+    const text = getTextContent(lastUser.content).toLowerCase();
+    const correctionSignals = [
+        'no ', 'wrong', 'that\'s not', 'thats not', 'incorrect', 'actually',
+        'you\'re wrong', 'youre wrong', 'that won\'t work', 'that wont work',
+        'not right', 'mistake', 'lied', 'lying', 'false', 'that\'s false',
+    ];
+    if (correctionSignals.some(sig => text.startsWith(sig) || text.includes(sig))) {
+        return getTextContent(lastUser.content);
+    }
+    return null;
+}
+// ─── Last Response Extraction ───────────────────────────────────────────────
+function extractLastExchange(messages) {
+    // Walk backwards to find the last user→assistant pair
+    let response = null;
+    let prompt = null;
+    for (let i = messages.length - 1; i >= 0; i--) {
+        const msg = messages[i];
+        if (!response && msg.role === 'assistant' && msg.content) {
+            response = getTextContent(msg.content);
+        }
+        else if (response && !prompt && msg.role === 'user') {
+            const text = getTextContent(msg.content);
+            // Skip internal planning injections
+            if (!text.startsWith('[Internal Planning')) {
+                prompt = text;
+                break;
+            }
+        }
+    }
+    if (!prompt || !response)
+        return null;
+    return { prompt, response };
+}
+// ─── Main Capture Function ──────────────────────────────────────────────────
+/**
+ * Capture an Ava interaction as a dataset example.
+ * Called after Agent.run() completes with the full message history.
+ * Fire-and-forget — never blocks the response.
+ */
+export async function captureInteraction(messages) {
+    try {
+        const config = await loadConfig();
+        if (!config.enabled)
+            return;
+        const mode = detectMode(messages);
+        if (!config.capture_modes.includes(mode))
+            return;
+        const exchange = extractLastExchange(messages);
+        if (!exchange)
+            return;
+        // Strip mode prefixes from prompt for cleaner training data
+        let cleanPrompt = exchange.prompt;
+        for (const [prefix] of MODE_PREFIXES) {
+            if (cleanPrompt.startsWith(prefix)) {
+                cleanPrompt = cleanPrompt.slice(prefix.length).trim();
+                break;
+            }
+        }
+        if (cleanPrompt.length < config.min_prompt_length)
+            return;
+        // Detect if this response was a correction of a previous one
+        const correction = detectCorrection(messages);
+        const quality = (correction && config.auto_label_corrections) ? 'sycophantic' : 'unlabelled';
+        const example = {
+            id: randomUUID(),
+            timestamp: new Date().toISOString(),
+            mode,
+            prompt: cleanPrompt,
+            response: exchange.response,
+            tools_used: extractToolsUsed(messages),
+            verified_before_answering: didVerifyBeforeAnswering(messages),
+            user_accepted: !correction,
+            correction,
+            quality,
+        };
+        // Ensure directory exists
+        await mkdir(RAW_DIR, { recursive: true });
+        // Append to daily JSONL file.
+        //
+        // UTC on purpose, and left that way in the 2026-08-17 local-day sweep.
+        // This is a PARTITION KEY, not a day anybody reads: files from machines in
+        // different timezones have to line up when the dataset is merged, and a
+        // local day would put two contributors' identical hour in different
+        // buckets. Nothing here is shown to a user, so there is no day to get
+        // wrong.
+        const date = new Date().toISOString().slice(0, 10);
+        const filePath = join(RAW_DIR, `${date}.jsonl`);
+        await appendFile(filePath, JSON.stringify(example) + '\n', 'utf-8');
+        logger.debug(`[dataset] Captured interaction: mode=${mode} quality=${quality} tools=${example.tools_used.length}`);
+    }
+    catch (error) {
+        // Never let dataset capture break the main flow
+        logger.debug(`[dataset] Capture failed: ${error instanceof Error ? error.message : String(error)}`);
+    }
+}
+//# sourceMappingURL=capture.js.map

package/dist/dataset/capture.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"capture.js","sourceRoot":"","sources":["../../src/dataset/capture.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AACjC,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AAC/D,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AACzC,OAAO,EAAE,QAAQ,EAAE,MAAM,sBAAsB,CAAC;AAChD,OAAO,EAAE,cAAc,EAAE,MAAM,kBAAkB,CAAC;AAElD,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AA0B3C,+EAA+E;AAE/E,MAAM,WAAW,GAAG,IAAI,CAAC,QAAQ,EAAE,UAAU,CAAC,CAAC;AAC/C,MAAM,OAAO,GAAG,IAAI,CAAC,WAAW,EAAE,KAAK,CAAC,CAAC;AACzC,MAAM,WAAW,GAAG,IAAI,CAAC,WAAW,EAAE,aAAa,CAAC,CAAC;AAErD,+EAA+E;AAE/E,MAAM,cAAc,GAAqB;IACvC,OAAO,EAAE,KAAK;IACd,aAAa,EAAE,CAAC,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,UAAU,EAAE,YAAY,CAAC;IAC1E,sBAAsB,EAAE,IAAI;IAC5B,iBAAiB,EAAE,EAAE;CACtB,CAAC;AAEF,IAAI,WAAW,GAA4B,IAAI,CAAC;AAChD,IAAI,cAAc,GAAG,CAAC,CAAC;AACvB,MAAM,UAAU,GAAG,MAAM,CAAC,CAAC,+BAA+B;AAE1D,KAAK,UAAU,UAAU;IACvB,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IACvB,IAAI,WAAW,IAAI,GAAG,GAAG,cAAc,GAAG,UAAU;QAAE,OAAO,WAAW,CAAC;IAEzE,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,WAAW,EAAE,OAAO,CAAC,CAAC;QACjD,WAAW,GAAG,EAAE,GAAG,cAAc,EAAE,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC;QACxD,cAAc,GAAG,GAAG,CAAC;QACrB,OAAO,WAAY,CAAC;IACtB,CAAC;IAAC,MAAM,CAAC;QACP,sDAAsD;QACtD,WAAW,GAAG,cAAc,CAAC;QAC7B,cAAc,GAAG,GAAG,CAAC;QACrB,OAAO,cAAc,CAAC;IACxB,CAAC;AACH,CAAC;AAED,+EAA+E;AAE/E,MAAM,aAAa,GAAwB;IACzC,CAAC,aAAa,EAAE,MAAM,CAAC;IACvB,CAAC,cAAc,EAAE,OAAO,CAAC;IACzB,CAAC,iBAAiB,EAAE,UAAU,CAAC;IAC/B,CAAC,mBAAmB,EAAE,YAAY,CAAC;IACnC,CAAC,aAAa,EAAE,MAAM,CAAC;CACxB,CAAC;AAEF,SAAS,UAAU,CAAC,QAAmB;IACrC,kFAAkF;IAClF,KAAK,IAAI,CAAC,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,GAAG,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;QACxB,IAAI,GAAG,CAAC,IAAI,KAAK,MAAM;YAAE,SAAS;QAClC,oCAAoC;QACpC,MAAM,IAAI,GAAG,cAAc,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;QACzC,IAAI,IAAI,CAAC,UAAU,CAAC,oBAAoB,CAAC;YAAE,SAAS;QAEpD,KAAK,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,aAAa,EAAE,CAAC;YAC3C,IAAI,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;gBAAE,OAAO,IAAI,CAAC;QAC3C,CAAC;QACD,kCAAkC;QAClC,OAAO,MAAM,CAAC;IAChB,CAAC;IACD,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,+EAA+E;AAE/E,SAAS,gBAAgB,CAAC,QAAmB;IAC3C,MAAM,KAAK,GAAG,IAAI,GAAG,EAAU,CAAC;IAChC,KAAK,MAAM,GAAG,IAAI,QAAQ,EAAE,CAAC;QAC3B,IAAI,GAAG,CAAC,IAAI,KAAK,WAAW,IAAI,YAAY,IAAI,GAAG,IAAI,GAAG,CAAC,UAAU,EAAE,CAAC;YACtE,KAAK,MAAM,EAAE,IAAI,GAAG,CAAC,UAAwB,EAAE,CAAC;gBAC9C,KAAK,CAAC,GAAG,CAAC,EAAE,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC;YAC9B,CAAC;QACH,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;AAC3B,CAAC;AAED,+EAA+E;AAE/E,8EAA8E;AAC9E,SAAS,wBAAwB,CAAC,QAAmB;IACnD,MAAM,iBAAiB,GAAG,IAAI,GAAG,CAAC;QAChC,WAAW,EAAE,MAAM,EAAE,MAAM,EAAE,gBAAgB,EAAE,aAAa;QAC5D,YAAY,EAAE,cAAc,EAAE,SAAS,EAAE,eAAe;QACxD,YAAY,EAAE,UAAU;KACzB,CAAC,CAAC;IAEH,MAAM,KAAK,GAAG,gBAAgB,CAAC,QAAQ,CAAC,CAAC;IACzC,OAAO,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,iBAAiB,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;AACnD,CAAC;AAED,+EAA+E;AAE/E,yFAAyF;AACzF,SAAS,gBAAgB,CAAC,QAAmB;IAC3C,sDAAsD;IACtD,IAAI,QAAQ,CAAC,MAAM,GAAG,CAAC;QAAE,OAAO,IAAI,CAAC;IAErC,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAC/C,MAAM,aAAa,GAAG,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACpD,MAAM,YAAY,GAAG,QAAQ,CAAC,QAAQ,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IAEnD,IAAI,QAAQ,EAAE,IAAI,KAAK,MAAM,IAAI,aAAa,EAAE,IAAI,KAAK,WAAW,IAAI,YAAY,EAAE,IAAI,KAAK,MAAM,EAAE,CAAC;QACtG,OAAO,IAAI,CAAC;IACd,CAAC;IAED,MAAM,IAAI,GAAG,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC,WAAW,EAAE,CAAC;IAC5D,MAAM,iBAAiB,GAAG;QACxB,KAAK,EAAE,OAAO,EAAE,aAAa,EAAE,WAAW,EAAE,WAAW,EAAE,UAAU;QACnE,eAAe,EAAE,aAAa,EAAE,kBAAkB,EAAE,gBAAgB;QACpE,WAAW,EAAE,SAAS,EAAE,MAAM,EAAE,OAAO,EAAE,OAAO,EAAE,eAAe;KAClE,CAAC;IAEF,IAAI,iBAAiB,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QAC9E,OAAO,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAC;IAC1C,CAAC;IAED,OAAO,IAAI,CAAC;AACd,CAAC;AAED,+EAA+E;AAE/E,SAAS,mBAAmB,CAAC,QAAmB;IAC9C,sDAAsD;IACtD,IAAI,QAAQ,GAAkB,IAAI,CAAC;IACnC,IAAI,MAAM,GAAkB,IAAI,CAAC;IAEjC,KAAK,IAAI,CAAC,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9C,MAAM,GAAG,GAAG,QAAQ,CAAC,CAAC,CAAC,CAAC;QACxB,IAAI,CAAC,QAAQ,IAAI,GAAG,CAAC,IAAI,KAAK,WAAW,IAAI,GAAG,CAAC,OAAO,EAAE,CAAC;YACzD,QAAQ,GAAG,cAAc,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;QACzC,CAAC;aAAM,IAAI,QAAQ,IAAI,CAAC,MAAM,IAAI,GAAG,CAAC,IAAI,KAAK,MAAM,EAAE,CAAC;YACtD,MAAM,IAAI,GAAG,cAAc,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC;YACzC,oCAAoC;YACpC,IAAI,CAAC,IAAI,CAAC,UAAU,CAAC,oBAAoB,CAAC,EAAE,CAAC;gBAC3C,MAAM,GAAG,IAAI,CAAC;gBACd,MAAM;YACR,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,CAAC,MAAM,IAAI,CAAC,QAAQ;QAAE,OAAO,IAAI,CAAC;IACtC,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,CAAC;AAC9B,CAAC;AAED,+EAA+E;AAE/E;;;;GAIG;AACH,MAAM,CAAC,KAAK,UAAU,kBAAkB,CAAC,QAAmB;IAC1D,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,MAAM,UAAU,EAAE,CAAC;QAClC,IAAI,CAAC,MAAM,CAAC,OAAO;YAAE,OAAO;QAE5B,MAAM,IAAI,GAAG,UAAU,CAAC,QAAQ,CAAC,CAAC;QAClC,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC,QAAQ,CAAC,IAAI,CAAC;YAAE,OAAO;QAEjD,MAAM,QAAQ,GAAG,mBAAmB,CAAC,QAAQ,CAAC,CAAC;QAC/C,IAAI,CAAC,QAAQ;YAAE,OAAO;QAEtB,4DAA4D;QAC5D,IAAI,WAAW,GAAG,QAAQ,CAAC,MAAM,CAAC;QAClC,KAAK,MAAM,CAAC,MAAM,CAAC,IAAI,aAAa,EAAE,CAAC;YACrC,IAAI,WAAW,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;gBACnC,WAAW,GAAG,WAAW,CAAC,KAAK,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,IAAI,EAAE,CAAC;gBACtD,MAAM;YACR,CAAC;QACH,CAAC;QAED,IAAI,WAAW,CAAC,MAAM,GAAG,MAAM,CAAC,iBAAiB;YAAE,OAAO;QAE1D,6DAA6D;QAC7D,MAAM,UAAU,GAAG,gBAAgB,CAAC,QAAQ,CAAC,CAAC;QAC9C,MAAM,OAAO,GAAG,CAAC,UAAU,IAAI,MAAM,CAAC,sBAAsB,CAAC,CAAC,CAAC,CAAC,aAAsB,CAAC,CAAC,CAAC,YAAqB,CAAC;QAE/G,MAAM,OAAO,GAAmB;YAC9B,EAAE,EAAE,UAAU,EAAE;YAChB,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;YACnC,IAAI;YACJ,MAAM,EAAE,WAAW;YACnB,QAAQ,EAAE,QAAQ,CAAC,QAAQ;YAC3B,UAAU,EAAE,gBAAgB,CAAC,QAAQ,CAAC;YACtC,yBAAyB,EAAE,wBAAwB,CAAC,QAAQ,CAAC;YAC7D,aAAa,EAAE,CAAC,UAAU;YAC1B,UAAU;YACV,OAAO;SACR,CAAC;QAEF,0BAA0B;QAC1B,MAAM,KAAK,CAAC,OAAO,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;QAE1C,8BAA8B;QAC9B,EAAE;QACF,uEAAuE;QACvE,2EAA2E;QAC3E,wEAAwE;QACxE,oEAAoE;QACpE,sEAAsE;QACtE,SAAS;QACT,MAAM,IAAI,GAAG,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC;QACnD,MAAM,QAAQ,GAAG,IAAI,CAAC,OAAO,EAAE,GAAG,IAAI,QAAQ,CAAC,CAAC;QAChD,MAAM,UAAU,CAAC,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,OAAO,CAAC,GAAG,IAAI,EAAE,OAAO,CAAC,CAAC;QAEpE,MAAM,CAAC,KAAK,CAAC,wCAAwC,IAAI,YAAY,OAAO,UAAU,OAAO,CAAC,UAAU,CAAC,MAAM,EAAE,CAAC,CAAC;IACrH,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACf,gDAAgD;QAChD,MAAM,CAAC,KAAK,CAAC,6BAA6B,KAAK,YAAY,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,EAAE,CAAC,CAAC;IACtG,CAAC;AACH,CAAC"}

package/dist/dataset/config.d.ts

@@ -0,0 +1,33 @@
+/**
+ * Dataset capture configuration.
+ *
+ * Lives at `<datasets-dir>/config.json`. Every knob is granular and
+ * defaults OFF — there is no way to accidentally opt someone in.
+ *
+ * Schema:
+ *   enabled              — master switch. If false, nothing is captured.
+ *   capture_modes        — which modes emit to datasets. Empty = none.
+ *   capture_datasets     — which of the 10 datasets get written. Empty = none.
+ *   redact_patterns      — extra regex strings applied at the write boundary
+ *                          on top of the built-in redactor defaults.
+ *   min_trajectory_length — skip trajectories with fewer than N events.
+ *                          Keeps noise (single-turn "hi" exchanges) out
+ *                          of the training data.
+ */
+import type { AvaMode } from './events.js';
+import type { DatasetName } from './routing.js';
+export interface DatasetConfig {
+    enabled: boolean;
+    capture_modes: AvaMode[];
+    capture_datasets: DatasetName[];
+    redact_patterns: string[];
+    min_trajectory_length: number;
+}
+export declare const DEFAULT_CONFIG: DatasetConfig;
+/** For tests and settings-UI writes. */
+export declare function invalidateConfigCache(): void;
+export declare function loadDatasetConfig(configPath: string): Promise<DatasetConfig>;
+export declare function saveDatasetConfig(configPath: string, config: DatasetConfig): Promise<void>;
+/** Convenience helper: standard path under a datasets directory. */
+export declare function configPathFor(datasetsDir: string): string;
+//# sourceMappingURL=config.d.ts.map

package/dist/dataset/config.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"config.d.ts","sourceRoot":"","sources":["../../src/dataset/config.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAIH,OAAO,KAAK,EAAE,OAAO,EAAE,MAAM,aAAa,CAAC;AAC3C,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAEhD,MAAM,WAAW,aAAa;IAC5B,OAAO,EAAE,OAAO,CAAC;IACjB,aAAa,EAAE,OAAO,EAAE,CAAC;IACzB,gBAAgB,EAAE,WAAW,EAAE,CAAC;IAChC,eAAe,EAAE,MAAM,EAAE,CAAC;IAC1B,qBAAqB,EAAE,MAAM,CAAC;CAC/B;AAED,eAAO,MAAM,cAAc,EAAE,aAM5B,CAAC;AAaF,wCAAwC;AACxC,wBAAgB,qBAAqB,IAAI,IAAI,CAE5C;AAID,wBAAsB,iBAAiB,CAAC,UAAU,EAAE,MAAM,GAAG,OAAO,CAAC,aAAa,CAAC,CAiBlF;AAED,wBAAsB,iBAAiB,CAAC,UAAU,EAAE,MAAM,EAAE,MAAM,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC,CAIhG;AAED,oEAAoE;AACpE,wBAAgB,aAAa,CAAC,WAAW,EAAE,MAAM,GAAG,MAAM,CAEzD"}

package/dist/dataset/config.js

@@ -0,0 +1,60 @@
+/**
+ * Dataset capture configuration.
+ *
+ * Lives at `<datasets-dir>/config.json`. Every knob is granular and
+ * defaults OFF — there is no way to accidentally opt someone in.
+ *
+ * Schema:
+ *   enabled              — master switch. If false, nothing is captured.
+ *   capture_modes        — which modes emit to datasets. Empty = none.
+ *   capture_datasets     — which of the 10 datasets get written. Empty = none.
+ *   redact_patterns      — extra regex strings applied at the write boundary
+ *                          on top of the built-in redactor defaults.
+ *   min_trajectory_length — skip trajectories with fewer than N events.
+ *                          Keeps noise (single-turn "hi" exchanges) out
+ *                          of the training data.
+ */
+import { join, dirname } from 'node:path';
+import { mkdir, readFile, writeFile } from 'node:fs/promises';
+export const DEFAULT_CONFIG = {
+    enabled: false,
+    capture_modes: [],
+    capture_datasets: [],
+    redact_patterns: [],
+    min_trajectory_length: 1,
+};
+const CACHE_TTL_MS = 30_000;
+let cache = null;
+/** For tests and settings-UI writes. */
+export function invalidateConfigCache() {
+    cache = null;
+}
+// ─── Load / save ────────────────────────────────────────────────────────────
+export async function loadDatasetConfig(configPath) {
+    const now = Date.now();
+    if (cache && cache.path === configPath && now - cache.readAt < CACHE_TTL_MS) {
+        return cache.config;
+    }
+    try {
+        const raw = await readFile(configPath, 'utf-8');
+        const parsed = JSON.parse(raw);
+        const config = { ...DEFAULT_CONFIG, ...parsed };
+        cache = { path: configPath, config, readAt: now };
+        return config;
+    }
+    catch {
+        // Missing or corrupt file — default to all-off.
+        cache = { path: configPath, config: DEFAULT_CONFIG, readAt: now };
+        return DEFAULT_CONFIG;
+    }
+}
+export async function saveDatasetConfig(configPath, config) {
+    await mkdir(dirname(configPath), { recursive: true });
+    await writeFile(configPath, JSON.stringify(config, null, 2), 'utf-8');
+    cache = { path: configPath, config, readAt: Date.now() };
+}
+/** Convenience helper: standard path under a datasets directory. */
+export function configPathFor(datasetsDir) {
+    return join(datasetsDir, 'config.json');
+}
+//# sourceMappingURL=config.js.map

package/dist/dataset/config.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"config.js","sourceRoot":"","sources":["../../src/dataset/config.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;GAeG;AAEH,OAAO,EAAE,IAAI,EAAE,OAAO,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,KAAK,EAAE,QAAQ,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AAY9D,MAAM,CAAC,MAAM,cAAc,GAAkB;IAC3C,OAAO,EAAE,KAAK;IACd,aAAa,EAAE,EAAE;IACjB,gBAAgB,EAAE,EAAE;IACpB,eAAe,EAAE,EAAE;IACnB,qBAAqB,EAAE,CAAC;CACzB,CAAC;AAUF,MAAM,YAAY,GAAG,MAAM,CAAC;AAC5B,IAAI,KAAK,GAAiB,IAAI,CAAC;AAE/B,wCAAwC;AACxC,MAAM,UAAU,qBAAqB;IACnC,KAAK,GAAG,IAAI,CAAC;AACf,CAAC;AAED,+EAA+E;AAE/E,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,UAAkB;IACxD,MAAM,GAAG,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IACvB,IAAI,KAAK,IAAI,KAAK,CAAC,IAAI,KAAK,UAAU,IAAI,GAAG,GAAG,KAAK,CAAC,MAAM,GAAG,YAAY,EAAE,CAAC;QAC5E,OAAO,KAAK,CAAC,MAAM,CAAC;IACtB,CAAC;IAED,IAAI,CAAC;QACH,MAAM,GAAG,GAAG,MAAM,QAAQ,CAAC,UAAU,EAAE,OAAO,CAAC,CAAC;QAChD,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAA2B,CAAC;QACzD,MAAM,MAAM,GAAkB,EAAE,GAAG,cAAc,EAAE,GAAG,MAAM,EAAE,CAAC;QAC/D,KAAK,GAAG,EAAE,IAAI,EAAE,UAAU,EAAE,MAAM,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;QAClD,OAAO,MAAM,CAAC;IAChB,CAAC;IAAC,MAAM,CAAC;QACP,gDAAgD;QAChD,KAAK,GAAG,EAAE,IAAI,EAAE,UAAU,EAAE,MAAM,EAAE,cAAc,EAAE,MAAM,EAAE,GAAG,EAAE,CAAC;QAClE,OAAO,cAAc,CAAC;IACxB,CAAC;AACH,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,iBAAiB,CAAC,UAAkB,EAAE,MAAqB;IAC/E,MAAM,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IACtD,MAAM,SAAS,CAAC,UAAU,EAAE,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,EAAE,OAAO,CAAC,CAAC;IACtE,KAAK,GAAG,EAAE,IAAI,EAAE,UAAU,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,CAAC,GAAG,EAAE,EAAE,CAAC;AAC3D,CAAC;AAED,oEAAoE;AACpE,MAAM,UAAU,aAAa,CAAC,WAAmB;IAC/C,OAAO,IAAI,CAAC,WAAW,EAAE,aAAa,CAAC,CAAC;AAC1C,CAAC"}

package/dist/dataset/consumer.d.ts

@@ -0,0 +1,33 @@
+/**
+ * Dataset capture consumer.
+ *
+ * Subscribes to every event on the `avaEvents` bus, applies the opt-in
+ * config gates, redacts payloads, and appends the event as a JSONL line
+ * to `<datasetsDir>/<dataset-name>/YYYY-MM-DD.jsonl`.
+ *
+ * Safety properties:
+ *   - Pure-functional from the agent's POV. We only read events the
+ *     emitter handed us (already structuredClone'd), never the live
+ *     conversation. The multi-writer corruption that disabled the
+ *     previous implementation is impossible here by construction.
+ *   - Per-file mutex serialises concurrent appends. Two agent
+ *     instances writing to the same daily file cannot interleave
+ *     half-lines.
+ *   - All errors are swallowed and logged — dataset capture failing
+ *     never takes down chat.
+ */
+/**
+ * Wait for every pending file write to finish. Test-only — production
+ * code never needs to block on capture.
+ */
+export declare function drainPendingWrites(): Promise<void>;
+export interface ConsumerOptions {
+    /** Override for tests. Defaults to ~/.ava/datasets */
+    datasetsDir?: string;
+}
+/**
+ * Subscribe the dataset consumer to the event bus. Idempotent — calling
+ * it twice is a no-op. Returns an unsubscribe function for tests.
+ */
+export declare function installDatasetConsumer(opts?: ConsumerOptions): () => void;
+//# sourceMappingURL=consumer.d.ts.map

package/dist/dataset/consumer.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"consumer.d.ts","sourceRoot":"","sources":["../../src/dataset/consumer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AA2BH;;;GAGG;AACH,wBAAsB,kBAAkB,IAAI,OAAO,CAAC,IAAI,CAAC,CAMxD;AAWD,MAAM,WAAW,eAAe;IAC9B,sDAAsD;IACtD,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAID;;;GAGG;AACH,wBAAgB,sBAAsB,CAAC,IAAI,GAAE,eAAoB,GAAG,MAAM,IAAI,CAiB7E"}

package/dist/dataset/consumer.js

@@ -0,0 +1,97 @@
+/**
+ * Dataset capture consumer.
+ *
+ * Subscribes to every event on the `avaEvents` bus, applies the opt-in
+ * config gates, redacts payloads, and appends the event as a JSONL line
+ * to `<datasetsDir>/<dataset-name>/YYYY-MM-DD.jsonl`.
+ *
+ * Safety properties:
+ *   - Pure-functional from the agent's POV. We only read events the
+ *     emitter handed us (already structuredClone'd), never the live
+ *     conversation. The multi-writer corruption that disabled the
+ *     previous implementation is impossible here by construction.
+ *   - Per-file mutex serialises concurrent appends. Two agent
+ *     instances writing to the same daily file cannot interleave
+ *     half-lines.
+ *   - All errors are swallowed and logged — dataset capture failing
+ *     never takes down chat.
+ */
+import { join } from 'node:path';
+import { mkdir, appendFile } from 'node:fs/promises';
+import { AVA_HOME } from '../core/constants.js';
+import { logger } from '../core/logger.js';
+import { avaEvents } from './emitter.js';
+import { eventToDataset } from './routing.js';
+import { loadDatasetConfig, configPathFor } from './config.js';
+import { redactValue } from './redactor.js';
+// ─── Per-file mutex ─────────────────────────────────────────────────────────
+const fileLocks = new Map();
+/**
+ * Serialise async work against a single path. Each call queues behind
+ * the previous one for the same path. Different paths run in parallel.
+ */
+async function withFileLock(path, fn) {
+    const prev = fileLocks.get(path) ?? Promise.resolve();
+    const next = prev.then(fn, fn); // run fn whether prev resolved or rejected
+    fileLocks.set(path, next.catch(() => undefined));
+    return next;
+}
+/**
+ * Wait for every pending file write to finish. Test-only — production
+ * code never needs to block on capture.
+ */
+export async function drainPendingWrites() {
+    // Snapshot the current chain heads, then await them. New writes that
+    // arrive after this snapshot are deliberately not awaited — caller
+    // controls ordering.
+    const chains = Array.from(fileLocks.values());
+    await Promise.allSettled(chains);
+}
+// ─── Path helpers ───────────────────────────────────────────────────────────
+function dailyFilePath(datasetsDir, datasetName, isoTimestamp) {
+    const date = isoTimestamp.slice(0, 10); // YYYY-MM-DD
+    return join(datasetsDir, datasetName, `${date}.jsonl`);
+}
+let installedUnsubscribe = null;
+/**
+ * Subscribe the dataset consumer to the event bus. Idempotent — calling
+ * it twice is a no-op. Returns an unsubscribe function for tests.
+ */
+export function installDatasetConsumer(opts = {}) {
+    if (installedUnsubscribe)
+        return installedUnsubscribe;
+    const datasetsDir = opts.datasetsDir ?? join(AVA_HOME, 'datasets');
+    const configPath = configPathFor(datasetsDir);
+    const off = avaEvents.onAll(async (event) => {
+        await handleEvent(event, datasetsDir, configPath);
+    });
+    installedUnsubscribe = () => {
+        off();
+        installedUnsubscribe = null;
+    };
+    logger.debug('[dataset] consumer installed');
+    return installedUnsubscribe;
+}
+// ─── Per-event handler ──────────────────────────────────────────────────────
+async function handleEvent(event, datasetsDir, configPath) {
+    try {
+        const config = await loadDatasetConfig(configPath);
+        if (!config.enabled)
+            return;
+        if (!config.capture_modes.includes(event.mode))
+            return;
+        const dataset = eventToDataset(event.event_type);
+        if (!config.capture_datasets.includes(dataset))
+            return;
+        const safe = redactValue(event, config.redact_patterns);
+        const filePath = dailyFilePath(datasetsDir, dataset, event.timestamp);
+        await withFileLock(filePath, async () => {
+            await mkdir(join(datasetsDir, dataset), { recursive: true });
+            await appendFile(filePath, JSON.stringify(safe) + '\n', 'utf-8');
+        });
+    }
+    catch (err) {
+        logger.debug(`[dataset] write failed: ${err instanceof Error ? err.message : String(err)}`);
+    }
+}
+//# sourceMappingURL=consumer.js.map

package/dist/dataset/consumer.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"consumer.js","sourceRoot":"","sources":["../../src/dataset/consumer.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;GAiBG;AAEH,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AACjC,OAAO,EAAE,KAAK,EAAE,UAAU,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,QAAQ,EAAE,MAAM,sBAAsB,CAAC;AAChD,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAC3C,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAAE,cAAc,EAAE,MAAM,cAAc,CAAC;AAC9C,OAAO,EAAE,iBAAiB,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAC/D,OAAO,EAAE,WAAW,EAAE,MAAM,eAAe,CAAC;AAG5C,+EAA+E;AAE/E,MAAM,SAAS,GAAG,IAAI,GAAG,EAA4B,CAAC;AAEtD;;;GAGG;AACH,KAAK,UAAU,YAAY,CAAI,IAAY,EAAE,EAAoB;IAC/D,MAAM,IAAI,GAAG,SAAS,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,OAAO,EAAE,CAAC;IACtD,MAAM,IAAI,GAAG,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC,2CAA2C;IAC3E,SAAS,CAAC,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC,CAAC;IACjD,OAAO,IAAI,CAAC;AACd,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,kBAAkB;IACtC,qEAAqE;IACrE,mEAAmE;IACnE,qBAAqB;IACrB,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,CAAC,CAAC;IAC9C,MAAM,OAAO,CAAC,UAAU,CAAC,MAAM,CAAC,CAAC;AACnC,CAAC;AAED,+EAA+E;AAE/E,SAAS,aAAa,CAAC,WAAmB,EAAE,WAAmB,EAAE,YAAoB;IACnF,MAAM,IAAI,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,aAAa;IACrD,OAAO,IAAI,CAAC,WAAW,EAAE,WAAW,EAAE,GAAG,IAAI,QAAQ,CAAC,CAAC;AACzD,CAAC;AASD,IAAI,oBAAoB,GAAwB,IAAI,CAAC;AAErD;;;GAGG;AACH,MAAM,UAAU,sBAAsB,CAAC,OAAwB,EAAE;IAC/D,IAAI,oBAAoB;QAAE,OAAO,oBAAoB,CAAC;IAEtD,MAAM,WAAW,GAAG,IAAI,CAAC,WAAW,IAAI,IAAI,CAAC,QAAQ,EAAE,UAAU,CAAC,CAAC;IACnE,MAAM,UAAU,GAAG,aAAa,CAAC,WAAW,CAAC,CAAC;IAE9C,MAAM,GAAG,GAAG,SAAS,CAAC,KAAK,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;QAC1C,MAAM,WAAW,CAAC,KAAK,EAAE,WAAW,EAAE,UAAU,CAAC,CAAC;IACpD,CAAC,CAAC,CAAC;IAEH,oBAAoB,GAAG,GAAG,EAAE;QAC1B,GAAG,EAAE,CAAC;QACN,oBAAoB,GAAG,IAAI,CAAC;IAC9B,CAAC,CAAC;IAEF,MAAM,CAAC,KAAK,CAAC,8BAA8B,CAAC,CAAC;IAC7C,OAAO,oBAAoB,CAAC;AAC9B,CAAC;AAED,+EAA+E;AAE/E,KAAK,UAAU,WAAW,CACxB,KAAe,EACf,WAAmB,EACnB,UAAkB;IAElB,IAAI,CAAC;QACH,MAAM,MAAM,GAAG,MAAM,iBAAiB,CAAC,UAAU,CAAC,CAAC;QACnD,IAAI,CAAC,MAAM,CAAC,OAAO;YAAE,OAAO;QAC5B,IAAI,CAAC,MAAM,CAAC,aAAa,CAAC,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC;YAAE,OAAO;QAEvD,MAAM,OAAO,GAAG,cAAc,CAAC,KAAK,CAAC,UAAU,CAAC,CAAC;QACjD,IAAI,CAAC,MAAM,CAAC,gBAAgB,CAAC,QAAQ,CAAC,OAAO,CAAC;YAAE,OAAO;QAEvD,MAAM,IAAI,GAAG,WAAW,CAAC,KAAK,EAAE,MAAM,CAAC,eAAe,CAAC,CAAC;QACxD,MAAM,QAAQ,GAAG,aAAa,CAAC,WAAW,EAAE,OAAO,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC;QAEtE,MAAM,YAAY,CAAC,QAAQ,EAAE,KAAK,IAAI,EAAE;YACtC,MAAM,KAAK,CAAC,IAAI,CAAC,WAAW,EAAE,OAAO,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;YAC7D,MAAM,UAAU,CAAC,QAAQ,EAAE,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,IAAI,EAAE,OAAO,CAAC,CAAC;QACnE,CAAC,CAAC,CAAC;IACL,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,MAAM,CAAC,KAAK,CACV,2BAA2B,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAC9E,CAAC;IACJ,CAAC;AACH,CAAC"}

package/dist/dataset/emitter.d.ts

@@ -0,0 +1,153 @@
+/**
+ * Typed Ava event emitter.
+ *
+ * Architecture:
+ *   - The emitter holds no business logic. It exists so the agent can
+ *     fire decision events without knowing what (if anything) is
+ *     listening. Capture is the consumer's problem; the agent just
+ *     emits.
+ *   - Trajectory context (trajectory_id, session_id, surface, mode,
+ *     model_id) lives in AsyncLocalStorage so emit sites don't have to
+ *     thread it through every call. Whoever opens an Agent.run() wraps
+ *     the run in `withTrajectory()` and every emit underneath inherits
+ *     the envelope.
+ *   - Payloads are deep-cloned at emit time via `structuredClone`. A
+ *     consumer that runs later (microtask, async I/O, whatever) can
+ *     never observe a mutated event because the mutation happened
+ *     after the snapshot. This is the root-cause fix for the
+ *     multi-writer corruption that disabled capture in the first
+ *     place.
+ *   - Handlers run on a microtask so emit() is non-blocking and can
+ *     never delay the user-visible response. A throwing handler
+ *     cannot crash the agent — errors are swallowed by design,
+ *     because dataset capture failing is never worth taking the chat
+ *     down for.
+ */
+import type { AvaEvent, AvaEventType, AvaEventPayloadMap, AvaSurface, AvaMode } from './events.js';
+export interface TrajectoryContext {
+    trajectory_id: string;
+    parent_trajectory_id?: string;
+    session_id: string;
+    surface: AvaSurface;
+    mode: AvaMode;
+    model_id: string;
+    /**
+     * Mutable list of tool names called so far in this trajectory. Used
+     * to populate `tool_choice.prev_tools_in_trajectory` without storing
+     * tool-specific state on the emitter. Mutated at emit time by the
+     * caller; safe because AsyncLocalStorage scopes the array to the
+     * trajectory's async chain.
+     */
+    toolsSoFar: string[];
+    /** When the trajectory opened — for total_duration_ms in tool_chain_complete. */
+    startedAt: number;
+    /**
+     * Outcome category for tool_chain_complete. Mutated by the agent when
+     * a non-default exit happens (loop limit, asked-user, gave-up). The
+     * trajectory wrapper reads this in its finally block and falls back
+     * to 'task_completed' if nothing set it.
+     */
+    outcome?: 'task_completed' | 'gave_up' | 'asked_user' | 'hit_loop_limit';
+    /**
+     * If the previous tool call in this trajectory failed, the event_id
+     * of its `tool_error` event. The next tool_choice consumes this to
+     * emit a `recovery_action` linking the recovery move back to the
+     * specific failure it's responding to. Cleared after consumption.
+     */
+    pendingErrorEventId?: string;
+    /**
+     * If a continuation stall was detected and a nudge was fired during
+     * this trajectory, the stall's event_id. The trajectory wrapper's
+     * finally block emits `continuation_nudge_fired` with `recovered`
+     * derived from `nudgeRecovered` (defaults to true — if the run made
+     * it to the end without re-entering the closure-fallback path, the
+     * nudge succeeded).
+     */
+    pendingStallEventId?: string;
+    nudgeRecovered?: boolean;
+    /**
+     * Files mutated by file_write / file_edit since the last verify_change run.
+     * Drives the post-edit verify hook in agent.ts: when this set is non-empty
+     * at closure time, run verify_change against the listed files. Cleared
+     * when verify runs (regardless of pass/fail).
+     *
+     * Lives on TrajectoryContext rather than agent state so a sub-agent / persona
+     * spawned inside the same trajectory inherits the unverified-files set
+     * automatically — useful when the orchestrator dispatches a Builder mid-task.
+     */
+    pendingVerifyFiles?: Set<string>;
+    /**
+     * Files that have passed verify_change since their last edit. Mostly
+     * informational — used to decide whether a closure attempt is OK
+     * (no pending files) or needs a verify nudge (pending files exist).
+     */
+    verifiedFiles?: Set<string>;
+    /**
+     * Timeline of verify_change failures during this trajectory. Each entry
+     * stores a stable signature derived from the failing checkId + first-error
+     * line so retries against the same root cause collide on the same signature
+     * even as the rest of the report shifts. Read by the error-loop detector
+     * to decide when to spawn a fresh-eyes pass.
+     */
+    verifyFailureHistory?: Array<{
+        signature: string;
+        check: string;
+        firedAt: number;
+    }>;
+    /**
+     * Set true once a fresh-eyes pass has been spawned for this trajectory's
+     * stuck verify failure. Prevents recursive escalation if the fresh-eyes
+     * recommendation is also broken — we surface to the user instead of
+     * spawning a third agent.
+     */
+    freshEyesEscalated?: boolean;
+}
+/**
+ * Open a trajectory context for the duration of `fn`. Every event
+ * emitted inside `fn` (sync or async) inherits this envelope. Wrap
+ * each top-level Agent.run() in this.
+ */
+export declare function withTrajectory<T>(ctx: Omit<TrajectoryContext, 'trajectory_id' | 'toolsSoFar' | 'startedAt'> & {
+    trajectory_id?: string;
+    toolsSoFar?: string[];
+    startedAt?: number;
+}, fn: () => T): T;
+/**
+ * Open a child trajectory whose `parent_trajectory_id` points at the
+ * currently-active trajectory. Use this when an orchestrator spawns a
+ * sub-agent (e.g. Builder inside a planning run) so we can reconstruct
+ * the orchestration tree at training time.
+ */
+export declare function withChildTrajectory<T>(overrides: Partial<TrajectoryContext>, fn: () => T): T;
+/** Read the current trajectory context. Returns undefined if not in a trajectory. */
+export declare function getTrajectory(): TrajectoryContext | undefined;
+type Handler<T extends AvaEventType> = (event: AvaEvent<T>) => void | Promise<void>;
+type AnyHandler = (event: AvaEvent) => void | Promise<void>;
+declare class AvaEventEmitter {
+    private handlers;
+    private wildcardHandlers;
+    /**
+     * Subscribe to one event type. Returns an unsubscribe function.
+     * Handlers run on a microtask after `emit` returns, so handler work
+     * cannot block or delay the emitter.
+     */
+    on<T extends AvaEventType>(type: T, handler: Handler<T>): () => void;
+    /** Subscribe to every event. Used by the dataset capture consumer. */
+    onAll(handler: AnyHandler): () => void;
+    /**
+     * Emit an event. Caller must be inside a `withTrajectory(...)` scope —
+     * if not, the emit is silently dropped (we don't throw because the
+     * agent shouldn't crash because of bookkeeping). Returns the
+     * generated event_id so callers can cross-reference it from a
+     * follow-up event (e.g. `tool_result.tool_choice_event_id`).
+     * Handlers run on a microtask, so emit() is non-blocking.
+     */
+    emit<T extends AvaEventType>(type: T, payload: AvaEventPayloadMap[T]): string;
+    /** For tests: drop every subscriber. */
+    clearAllHandlers(): void;
+    private runHandlerSafe;
+}
+/** Singleton — one event bus per process. */
+export declare const avaEvents: AvaEventEmitter;
+export {};
+//# sourceMappingURL=emitter.d.ts.map

package/dist/dataset/emitter.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"emitter.d.ts","sourceRoot":"","sources":["../../src/dataset/emitter.ts"],"names":[],"mappings":"AAAA;;;;;;;;;;;;;;;;;;;;;;;;GAwBG;AAIH,OAAO,KAAK,EACV,QAAQ,EACR,YAAY,EACZ,kBAAkB,EAClB,UAAU,EACV,OAAO,EACR,MAAM,aAAa,CAAC;AAIrB,MAAM,WAAW,iBAAiB;IAChC,aAAa,EAAE,MAAM,CAAC;IACtB,oBAAoB,CAAC,EAAE,MAAM,CAAC;IAC9B,UAAU,EAAE,MAAM,CAAC;IACnB,OAAO,EAAE,UAAU,CAAC;IACpB,IAAI,EAAE,OAAO,CAAC;IACd,QAAQ,EAAE,MAAM,CAAC;IACjB;;;;;;OAMG;IACH,UAAU,EAAE,MAAM,EAAE,CAAC;IACrB,iFAAiF;IACjF,SAAS,EAAE,MAAM,CAAC;IAClB;;;;;OAKG;IACH,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,GAAG,YAAY,GAAG,gBAAgB,CAAC;IACzE;;;;;OAKG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAC7B;;;;;;;OAOG;IACH,mBAAmB,CAAC,EAAE,MAAM,CAAC;IAC7B,cAAc,CAAC,EAAE,OAAO,CAAC;IACzB;;;;;;;;;OASG;IACH,kBAAkB,CAAC,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IACjC;;;;OAIG;IACH,aAAa,CAAC,EAAE,GAAG,CAAC,MAAM,CAAC,CAAC;IAC5B;;;;;;OAMG;IACH,oBAAoB,CAAC,EAAE,KAAK,CAAC;QAAE,SAAS,EAAE,MAAM,CAAC;QAAC,KAAK,EAAE,MAAM,CAAC;QAAC,OAAO,EAAE,MAAM,CAAA;KAAE,CAAC,CAAC;IACpF;;;;;OAKG;IACH,kBAAkB,CAAC,EAAE,OAAO,CAAC;CAC9B;AAID;;;;GAIG;AACH,wBAAgB,cAAc,CAAC,CAAC,EAC9B,GAAG,EAAE,IAAI,CAAC,iBAAiB,EAAE,eAAe,GAAG,YAAY,GAAG,WAAW,CAAC,GAAG;IAC3E,aAAa,CAAC,EAAE,MAAM,CAAC;IACvB,UAAU,CAAC,EAAE,MAAM,EAAE,CAAC;IACtB,SAAS,CAAC,EAAE,MAAM,CAAC;CACpB,EACD,EAAE,EAAE,MAAM,CAAC,GACV,CAAC,CAKH;AAED;;;;;GAKG;AACH,wBAAgB,mBAAmB,CAAC,CAAC,EACnC,SAAS,EAAE,OAAO,CAAC,iBAAiB,CAAC,EACrC,EAAE,EAAE,MAAM,CAAC,GACV,CAAC,CA2BH;AAED,qFAAqF;AACrF,wBAAgB,aAAa,IAAI,iBAAiB,GAAG,SAAS,CAE7D;AAID,KAAK,OAAO,CAAC,CAAC,SAAS,YAAY,IAAI,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC,KAAK,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;AACpF,KAAK,UAAU,GAAG,CAAC,KAAK,EAAE,QAAQ,KAAK,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC,CAAC;AAE5D,cAAM,eAAe;IACnB,OAAO,CAAC,QAAQ,CAAuD;IACvE,OAAO,CAAC,gBAAgB,CAAyB;IAEjD;;;;OAIG;IACH,EAAE,CAAC,CAAC,SAAS,YAAY,EAAE,IAAI,EAAE,CAAC,EAAE,OAAO,EAAE,OAAO,CAAC,CAAC,CAAC,GAAG,MAAM,IAAI;IAYpE,sEAAsE;IACtE,KAAK,CAAC,OAAO,EAAE,UAAU,GAAG,MAAM,IAAI;IAOtC;;;;;;;OAOG;IACH,IAAI,CAAC,CAAC,SAAS,YAAY,EAAE,IAAI,EAAE,CAAC,EAAE,OAAO,EAAE,kBAAkB,CAAC,CAAC,CAAC,GAAG,MAAM;IAkC7E,wCAAwC;IACxC,gBAAgB,IAAI,IAAI;IAKxB,OAAO,CAAC,cAAc;CAYvB;AAED,6CAA6C;AAC7C,eAAO,MAAM,SAAS,iBAAwB,CAAC"}