@gobing-ai/ts-llm-jsonl-importer 0.4.48 → 0.4.50

package/README.md

@@ -7,7 +7,7 @@ Generic JSONL import pipeline for AI-agent history-style files: discover files,
 `ts-llm-jsonl-importer` is a common JSONL importer. It is intentionally not a conversation-history domain model. Built-in source definitions cover common agent file shapes, but downstream systems consume normalized ETL rows and ledger metadata.
 
 | Export | Purpose |
-|--------|---------|
+| -------- | --------- |
 | `runJsonlImport()` | Runs discovery, parsing, validation, redaction, dedupe, and persistence. Accepts a built-in source string **or** a custom `SourceDefinition`. |
 | `applyHistoryImportSchema()` | Installs importer-owned checkpoint, ledger, and typed contract tables |
 | `SOURCE_DEFINITIONS` | Built-in source definitions |
@@ -19,7 +19,7 @@ Generic JSONL import pipeline for AI-agent history-style files: discover files,
 | `sha256()` / `stableJson()` | Stable hash helpers used by the ledger |
 | `HISTORY_IMPORT_SCHEMA_SQL` | SQL schema string for explicit migration flows |
 
-Built-in source keys are `claude`, `codex`, `gemini`, `pi`, `opencode`, `antigravity`, `openclaw`, `omp`, `grok`, and `agy`.
+Built-in source keys are `claude`, `codex`, `gemini`, `pi`, `opencode`, `antigravity`, `openclaw`, `omp`, `grok`, and `agy`. The `agy` source scans `~/.gemini/antigravity-cli`, covering both `history.jsonl` and `brain/**` transcripts; conversation `.db` stores under `conversations/` are not imported.
 
 ## Installation
 
@@ -52,7 +52,7 @@ result.skippedDuplicates;
 ## Import Modes
 
 | Mode | Behavior |
-|------|----------|
+| ------ | ---------- |
 | `incremental` | Reads each file from the last imported line checkpoint |
 | `full` | Clears checkpoints for selected files, scans all lines, and still deduplicates by ledger hash |
 | `force-file` | Scans selected files without using the checkpoint, while preserving ledger-based duplicate protection |
@@ -124,7 +124,7 @@ Use `parseErrors` for invalid JSON or non-object rows. Use `validationErrors` fo
 The schema contains:
 
 | Table | Purpose |
-|-------|---------|
+| ------- | --------- |
 | `history_import_checkpoint` | Per-source/per-file last imported line |
 | `history_import_ledger` | Stable hash ledger for dedupe and provenance |
 | `history_message` | Typed normalized messages from built-in forensic mappers |

package/dist/mappers.d.ts

@@ -3,10 +3,11 @@ import type { JsonObject, SplitEntry, TransformContext } from './types';
 /** Compute args_digest: sha256 of stable-jsonified, key-sorted, redacted args. */
 export declare function argsDigest(args: unknown): string;
 /**
- * Return JSON-stringified raw args when the tool is a todo-writing tool for its source,
- * otherwise undefined. Codex `arguments` arrives as a JSON string already — store as-is.
+ * Return retained raw args for a typed tool call: the shell command string for
+ * recognized bash/shell/exec/command tools, sanitized valid JSON otherwise, and
+ * undefined when the event carries no invocation arguments.
  */
-export declare function maybeArgsRaw(source: string, toolName: string, args: unknown): string | undefined;
+export declare function maybeArgsRaw(_source: string, toolName: string, args: unknown): string | undefined;
 /** Map a Claude Code JSONL record into one or more ETL split entries. */
 export declare function claudeSplit(raw: Record<string, unknown>): readonly SplitEntry[];
 /** Map a Pi JSONL record into one or more ETL split entries. */

package/dist/mappers.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"mappers.d.ts","sourceRoot":"","sources":["../src/mappers.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,OAAO,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAC;AAuExE,kFAAkF;AAClF,wBAAgB,UAAU,CAAC,IAAI,EAAE,OAAO,GAAG,MAAM,CAGhD;AA6CD;;;GAGG;AACH,wBAAgB,YAAY,CAAC,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CAMhG;AAMD,yEAAyE;AACzE,wBAAgB,WAAW,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,SAAS,UAAU,EAAE,CAkH/E;AAMD,gEAAgE;AAChE,wBAAgB,OAAO,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CAmHvG;AAmBD,kEAAkE;AAClE,wBAAgB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CA+GxG;AAoBD,mFAAmF;AACnF,MAAM,WAAW,mBAAmB;IAChC,yEAAyE;IACzE,UAAU,EAAE,MAAM,CAAC;IACnB,wEAAwE;IACxE,UAAU,EAAE,MAAM,GAAG,SAAS,CAAC;IAC/B,yDAAyD;IACzD,WAAW,EAAE,MAAM,GAAG,SAAS,CAAC;CACnC;AAED,+FAA+F;AAC/F,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CASrE;AAED;;;;;;;GAOG;AACH,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,mBAAmB,GAAG,IAAI,CAS5F;AAMD,wEAAwE;AACxE,MAAM,WAAW,sBAAsB;IACnC,uFAAuF;IACvF,UAAU,EAAE,MAAM,CAAC;IACnB,yDAAyD;IACzD,WAAW,EAAE,MAAM,GAAG,SAAS,CAAC;IAChC,wDAAwD;IACxD,UAAU,EAAE,MAAM,GAAG,SAAS,CAAC;IAC/B,2DAA2D;IAC3D,WAAW,EAAE,MAAM,CAAC;CACvB;AAED;;;;;;;;GAQG;AACH,wBAAgB,sBAAsB,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,sBAAsB,GAAG,IAAI,CA8BlG;AAMD,mEAAmE;AACnE,wBAAgB,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CA4L1G;AAMD,gFAAgF;AAChF,wBAAgB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CAmHxG;AAMD,gGAAgG;AAChG,wBAAgB,WAAW,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CAiF3G;AAMD;;;;;;GAMG;AACH,wBAAgB,uBAAuB,CAAC,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAgB9F;AAiMD,kEAAkE;AAClE,wBAAgB,SAAS,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,SAAS,UAAU,EAAE,CA8S7E;AAgFD,wGAAwG;AACxG,wBAAgB,gBAAgB,CAAC,GAAG,EAAE,UAAU,GAAG,MAAM,CAKxD;AAMD,gDAAgD;AAChD,eAAO,MAAM,gBAAgB,wBAAsE,CAAC;AACpG,4CAA4C;AAC5C,eAAO,MAAM,YAAY,wBAAsE,CAAC;AAChG,6CAA6C;AAC7C,eAAO,MAAM,aAAa,wBAAsE,CAAC;AACjG,+CAA+C;AAC/C,eAAO,MAAM,eAAe,wBAAsE,CAAC;AACnG,6CAA6C;AAC7C,eAAO,MAAM,aAAa,wBAAsE,CAAC;AACjG,8CAA8C;AAC9C,eAAO,MAAM,cAAc,wBAAsE,CAAC;AAClG,gDAAgD;AAChD,eAAO,MAAM,gBAAgB,wBAAsE,CAAC;AAQpG,oDAAoD;AACpD,eAAO,MAAM,aAAa,gCAAoB,CAAC;AAC/C,gDAAgD;AAChD,eAAO,MAAM,SAAS,gCAAoB,CAAC;AAC3C,iDAAiD;AACjD,eAAO,MAAM,UAAU,gCAAoB,CAAC;AAC5C,mDAAmD;AACnD,eAAO,MAAM,YAAY,gCAAoB,CAAC;AAC9C,iDAAiD;AACjD,eAAO,MAAM,UAAU,gCAAoB,CAAC;AAC5C,kDAAkD;AAClD,eAAO,MAAM,WAAW,gCAAoB,CAAC;AAC7C,oDAAoD;AACpD,eAAO,MAAM,aAAa,gCAAoB,CAAC"}
+{"version":3,"file":"mappers.d.ts","sourceRoot":"","sources":["../src/mappers.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,OAAO,KAAK,EAAE,UAAU,EAAE,UAAU,EAAE,gBAAgB,EAAE,MAAM,SAAS,CAAC;AAuExE,kFAAkF;AAClF,wBAAgB,UAAU,CAAC,IAAI,EAAE,OAAO,GAAG,MAAM,CAGhD;AAwHD;;;;GAIG;AACH,wBAAgB,YAAY,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CAwBjG;AAMD,yEAAyE;AACzE,wBAAgB,WAAW,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,SAAS,UAAU,EAAE,CAkH/E;AAMD,gEAAgE;AAChE,wBAAgB,OAAO,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CAmHvG;AAmBD,kEAAkE;AAClE,wBAAgB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CA+GxG;AAoBD,mFAAmF;AACnF,MAAM,WAAW,mBAAmB;IAChC,yEAAyE;IACzE,UAAU,EAAE,MAAM,CAAC;IACnB,wEAAwE;IACxE,UAAU,EAAE,MAAM,GAAG,SAAS,CAAC;IAC/B,yDAAyD;IACzD,WAAW,EAAE,MAAM,GAAG,SAAS,CAAC;CACnC;AAED,+FAA+F;AAC/F,wBAAgB,kBAAkB,CAAC,KAAK,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CASrE;AAED;;;;;;;GAOG;AACH,wBAAgB,mBAAmB,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,mBAAmB,GAAG,IAAI,CAS5F;AAMD,wEAAwE;AACxE,MAAM,WAAW,sBAAsB;IACnC,uFAAuF;IACvF,UAAU,EAAE,MAAM,CAAC;IACnB,yDAAyD;IACzD,WAAW,EAAE,MAAM,GAAG,SAAS,CAAC;IAChC,wDAAwD;IACxD,UAAU,EAAE,MAAM,GAAG,SAAS,CAAC;IAC/B,2DAA2D;IAC3D,WAAW,EAAE,MAAM,CAAC;CACvB;AAED;;;;;;;;GAQG;AACH,wBAAgB,sBAAsB,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,sBAAsB,GAAG,IAAI,CA8BlG;AAMD,mEAAmE;AACnE,wBAAgB,UAAU,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CA4L1G;AAMD,gFAAgF;AAChF,wBAAgB,QAAQ,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CAmIxG;AAMD,gGAAgG;AAChG,wBAAgB,WAAW,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EAAE,OAAO,CAAC,EAAE,gBAAgB,GAAG,SAAS,UAAU,EAAE,CAiF3G;AAMD;;;;;;GAMG;AACH,wBAAgB,uBAAuB,CAAC,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,MAAM,GAAG,MAAM,GAAG,SAAS,CAgB9F;AAiMD,kEAAkE;AAClE,wBAAgB,SAAS,CAAC,GAAG,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,SAAS,UAAU,EAAE,CA8S7E;AAgFD,wGAAwG;AACxG,wBAAgB,gBAAgB,CAAC,GAAG,EAAE,UAAU,GAAG,MAAM,CAKxD;AAMD,gDAAgD;AAChD,eAAO,MAAM,gBAAgB,wBAAsE,CAAC;AACpG,4CAA4C;AAC5C,eAAO,MAAM,YAAY,wBAAsE,CAAC;AAChG,6CAA6C;AAC7C,eAAO,MAAM,aAAa,wBAAsE,CAAC;AACjG,+CAA+C;AAC/C,eAAO,MAAM,eAAe,wBAAsE,CAAC;AACnG,6CAA6C;AAC7C,eAAO,MAAM,aAAa,wBAAsE,CAAC;AACjG,8CAA8C;AAC9C,eAAO,MAAM,cAAc,wBAAsE,CAAC;AAClG,gDAAgD;AAChD,eAAO,MAAM,gBAAgB,wBAAsE,CAAC;AAQpG,oDAAoD;AACpD,eAAO,MAAM,aAAa,gCAAoB,CAAC;AAC/C,gDAAgD;AAChD,eAAO,MAAM,SAAS,gCAAoB,CAAC;AAC3C,iDAAiD;AACjD,eAAO,MAAM,UAAU,gCAAoB,CAAC;AAC5C,mDAAmD;AACnD,eAAO,MAAM,YAAY,gCAAoB,CAAC;AAC9C,iDAAiD;AACjD,eAAO,MAAM,UAAU,gCAAoB,CAAC;AAC5C,kDAAkD;AAClD,eAAO,MAAM,WAAW,gCAAoB,CAAC;AAC7C,oDAAoD;AACpD,eAAO,MAAM,aAAa,gCAAoB,CAAC"}

package/dist/mappers.js

@@ -91,37 +91,122 @@ function redactArgs(args) {
     return args;
 }
 /**
- * Per-source todo-writing tool names whose raw arguments are retained for phase detection.
- *
- * Evidence: 0489 R4 confirmed omp/pi/claude; codex/grok/agy probed task 0553 R3 from real
- * session JSONL. agy has no on-disk session format (VS Code fork); gemini deferred by the
- * 2026-08-06 source-support ruling. Task 0578 R3: omp `todo_write`, pi `manage_todo_list`,
- * opencode `todowrite`/`todoread` measured live in .spur/spur.db (3,237/21, 1,043, 379/2
- * rows with NULL args_raw).
+ * Retained tool-call arguments: recognized shell tools keep the extracted plain
+ * command (attribution compatibility, capped); every other typed tool call keeps
+ * sanitized-but-valid JSON. `args_digest` is computed elsewhere from the unpruned
+ * args and is unaffected; secret redaction stays at the persistence `redactRecord`
+ * seam (task 0064 R5).
  */
-const TODO_TOOL_ALLOWLIST = {
-    claude: ['TodoWrite'],
-    opencode: ['todowrite', 'todoread'],
-    codex: ['update_plan'],
-    omp: ['TodoWrite', 'todo', 'todo_write'],
-    grok: ['todo_write'],
-    pi: ['todo', 'manage_todo_list'],
-    agy: [],
-    gemini: [],
-};
+const BASH_COMMAND_MAX_CHARS = 8192;
+const GENERIC_ARGS_MAX_CHARS = 8192;
+/** Extract the shell command from a bash tool's raw input; undefined when absent. */
+function bashCommandOf(args) {
+    if (args === null || typeof args !== 'object')
+        return undefined;
+    const rec = args;
+    const command = rec.command ?? rec.CommandLine ?? rec.cmd ?? rec.script;
+    if (typeof command === 'string')
+        return command;
+    if (Array.isArray(command))
+        return command.map(String).join(' ');
+    return undefined;
+}
+/** Bulky payload keys (normalized) pruned to an omission marker; camelCase and snake_case share one rule. */
+const BULKY_PAYLOAD_KEYS = new Set([
+    'content',
+    'codecontent',
+    'replacementcontent',
+    'targetcontent',
+    'filedata',
+    'buffer',
+    'blob',
+    'image',
+    'imagebase64',
+]);
+/** Lowercase a key and drop separators so code_content and codeContent compare equal. */
+function normalizeKey(key) {
+    return key.toLowerCase().replace(/[^a-z0-9]/g, '');
+}
+/** Tool names whose command-shaped args retain the existing plain-string representation. */
+function isShellTool(toolName) {
+    return /^(?:bash|shell|cmd|command|exec|execcommand|executecommand|runcommand|shellcommand|runshellcommand)$/.test(normalizeKey(toolName));
+}
 /**
- * Return JSON-stringified raw args when the tool is a todo-writing tool for its source,
- * otherwise undefined. Codex `arguments` arrives as a JSON string already — store as-is.
+ * Sanitize any JSON value recursively through objects AND arrays: bulky payload
+ * strings under known keys become `[omitted N chars]`, other strings longer than
+ * 1,000 characters keep a prefix plus an explicit original-length marker, and
+ * every primitive and short payload is preserved untouched (task 0064 R2).
  */
-export function maybeArgsRaw(source, toolName, args) {
-    const allow = TODO_TOOL_ALLOWLIST[source];
-    if (allow === undefined || !allow.includes(toolName))
-        return undefined;
-    if (typeof args === 'string')
-        return args;
+function sanitizeValue(value) {
+    if (typeof value === 'string') {
+        return value.length > 1000 ? `${value.slice(0, 1000)}[truncated ${value.length} chars]` : value;
+    }
+    if (Array.isArray(value))
+        return value.map(sanitizeValue);
+    if (value !== null && typeof value === 'object') {
+        const out = {};
+        for (const [key, entry] of Object.entries(value)) {
+            out[key] =
+                typeof entry === 'string' && entry.length > 120 && BULKY_PAYLOAD_KEYS.has(normalizeKey(key))
+                    ? `[omitted ${entry.length} chars]`
+                    : sanitizeValue(entry);
+        }
+        return out;
+    }
+    // JSON leaves (number/boolean/null) pass through unchanged; JSON args never
+    // reach here as undefined/bigint/symbol/function.
+    return value;
+}
+/**
+ * Serialize sanitized args within the retention ceiling without ever slicing JSON
+ * text: values that still serialize past the ceiling collapse into a small valid
+ * JSON wrapper carrying `_truncated`, the omitted character count, and a bounded
+ * preview reduced until the wrapper itself fits (task 0064 R3/AC4).
+ */
+function serializeBounded(value) {
+    const serialized = JSON.stringify(sanitizeValue(value));
+    if (serialized.length <= GENERIC_ARGS_MAX_CHARS)
+        return serialized;
+    let preview = serialized;
+    for (;;) {
+        const wrapper = JSON.stringify({
+            _truncated: true,
+            omittedChars: serialized.length - preview.length,
+            preview,
+        });
+        if (wrapper.length <= GENERIC_ARGS_MAX_CHARS)
+            return wrapper;
+        preview = preview.slice(0, preview.length - (wrapper.length - GENERIC_ARGS_MAX_CHARS));
+    }
+}
+/**
+ * Return retained raw args for a typed tool call: the shell command string for
+ * recognized bash/shell/exec/command tools, sanitized valid JSON otherwise, and
+ * undefined when the event carries no invocation arguments.
+ */
+export function maybeArgsRaw(_source, toolName, args) {
     if (args === undefined || args === null)
         return undefined;
-    return JSON.stringify(args);
+    let value = args;
+    if (typeof args === 'string') {
+        // Pre-stringified arguments (Codex) re-enter the structured path so the
+        // retained value parses and prunes like native JSON.
+        try {
+            value = JSON.parse(args);
+        }
+        catch {
+            return args.length > BASH_COMMAND_MAX_CHARS ? args.slice(0, BASH_COMMAND_MAX_CHARS) : args;
+        }
+    }
+    if (value !== null && typeof value === 'object') {
+        const cmd = bashCommandOf(value);
+        if (cmd !== undefined && isShellTool(toolName)) {
+            return cmd.length > BASH_COMMAND_MAX_CHARS ? cmd.slice(0, BASH_COMMAND_MAX_CHARS) : cmd;
+        }
+        return serializeBounded(value);
+    }
+    // JSON primitives (number/boolean) retain as valid JSON scalars.
+    return JSON.stringify(value);
 }
 // ---------------------------------------------------------------------------
 // Claude mapper
@@ -748,67 +833,83 @@ export function codexSplit(raw, context) {
 export function agySplit(raw, context) {
     const entries = [];
     const recordType = String(raw.type ?? '');
-    const explicitSessionId = s(raw.session_id, raw.conversation_id);
+    const explicitSessionId = s(raw.session_id, raw.conversation_id, raw.conversationId);
     const pathSessionId = context?.sourceFile ? sessionIdFromSourcePath('agy', context.sourceFile) : undefined;
     const sessionId = explicitSessionId ?? pathSessionId ?? 'unknown';
-    const seq = typeof raw.seq === 'number' ? raw.seq : typeof raw.step_index === 'number' ? raw.step_index : 0;
+    const seq = typeof raw.seq === 'number'
+        ? raw.seq
+        : typeof raw.step_index === 'number'
+            ? raw.step_index
+            : (context?.sourceLine ?? 0);
     const ts = timestampOf(raw.created_at, raw.timestamp, raw.ts);
     let role;
     let disposition;
     let contentText = null;
-    // Classification from task 0463 field map (agy transcript types).
-    switch (recordType) {
-        case 'USER_INPUT':
-        case 'ASK_QUESTION':
-            role = 'user';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.text, raw.message) ?? null;
-            break;
-        case 'PLANNER_RESPONSE':
-            role = 'assistant';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.text, raw.message, raw.plan) ?? null;
-            break;
-        case 'ERROR_MESSAGE':
-            role = 'system';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.text, raw.message) ?? null;
-            break;
-        case 'RUN_COMMAND':
-        case 'VIEW_FILE':
-        case 'GREP_SEARCH':
-        case 'LIST_DIRECTORY':
-        case 'READ_URL_CONTENT':
-            role = 'tool';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.command, raw.file_path, raw.pattern) ?? null;
-            break;
-        case 'CODE_ACTION':
-        case 'INVOKE_SUBAGENT':
-            role = 'tool';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.text) ?? null;
-            break;
-        case 'CONVERSATION_HISTORY':
-        case 'CHECKPOINT':
-        case 'GENERIC':
-            role = 'meta';
-            disposition = 'meta';
-            contentText = s(raw.content, raw.text) ?? null;
-            break;
-        default:
-            if (recordType.length === 0) {
-                role = 'unknown';
-                disposition = 'unknown';
-            }
-            else {
-                // Known discriminator present but not in the frozen 0463 keep/meta table —
-                // still a determined type; treat as meta bookkeeping rather than unknown.
+    let messageRecordType = recordType;
+    // history.jsonl prompt index: {display, timestamp, workspace, conversationId?, type?}.
+    // One branch covers absent type, 'slash_command', 'shell', and any future producer type.
+    // `display` never appears on legacy brain transcripts, so this cannot collide with 0463.
+    if (raw.display !== undefined) {
+        role = 'user';
+        disposition = 'keep';
+        contentText = s(raw.display) ?? null;
+        messageRecordType = recordType.length > 0 ? recordType : 'USER_INPUT';
+    }
+    else {
+        // Classification from task 0463 field map (agy transcript types).
+        switch (recordType) {
+            case 'USER_INPUT':
+            case 'ASK_QUESTION':
+                role = 'user';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.text, raw.message) ?? null;
+                break;
+            case 'PLANNER_RESPONSE':
+                role = 'assistant';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.text, raw.message, raw.plan) ?? null;
+                break;
+            case 'ERROR_MESSAGE':
+                role = 'system';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.text, raw.message) ?? null;
+                break;
+            case 'RUN_COMMAND':
+            case 'VIEW_FILE':
+            case 'GREP_SEARCH':
+            case 'LIST_DIRECTORY':
+            case 'READ_URL_CONTENT':
+                role = 'tool';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.command, raw.file_path, raw.pattern) ?? null;
+                break;
+            case 'CODE_ACTION':
+            case 'INVOKE_SUBAGENT':
+                role = 'tool';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.text) ?? null;
+                break;
+            case 'CONVERSATION_HISTORY':
+            case 'CHECKPOINT':
+            case 'GENERIC':
                 role = 'meta';
                 disposition = 'meta';
                 contentText = s(raw.content, raw.text) ?? null;
-            }
-            break;
+                break;
+            default:
+                if (recordType.length === 0) {
+                    role = 'unknown';
+                    disposition = 'unknown';
+                }
+                else {
+                    // Known discriminator present but not in the frozen 0463 keep/meta table —
+                    // still a determined type; treat as meta bookkeeping rather than unknown.
+                    role = 'meta';
+                    disposition = 'meta';
+                    contentText = s(raw.content, raw.text) ?? null;
+                }
+                break;
+        }
     }
     const messageSplitIndex = entries.length;
     entries.push({
@@ -817,7 +918,7 @@ export function agySplit(raw, context) {
             session_id: sessionId,
             seq,
             role,
-            record_type: recordType,
+            record_type: messageRecordType,
             disposition,
             ts,
             duration_ms: undefined,
@@ -828,7 +929,7 @@ export function agySplit(raw, context) {
             cache_write_tokens: null,
             cost_usd: null,
             content_text: contentText,
-            cwd: null,
+            cwd: s(raw.workspace, raw.cwd) ?? null,
             provenance: 'ambient',
         },
     });

package/dist/sources.js

@@ -104,7 +104,7 @@ export const SOURCE_DEFINITIONS = {
     omp: customSourceDefinition('omp', 'OMP', ['.omp/agent/sessions'], ['*.jsonl'], ompSplit, OMP_FIELD_MAP, OMP_SCHEMA),
     grok: customSourceDefinition('grok', 'Grok', ['.grok/sessions'], ['*.jsonl'], grokSplit, GROK_FIELD_MAP, GROK_SCHEMA),
     agy: {
-        ...customSourceDefinition('agy', 'Antigravity CLI', ['.gemini/antigravity-cli/brain'], ['*.jsonl'], agySplit, AGY_FIELD_MAP, AGY_SCHEMA),
+        ...customSourceDefinition('agy', 'Antigravity CLI', ['.gemini/antigravity-cli'], ['*.jsonl'], agySplit, AGY_FIELD_MAP, AGY_SCHEMA),
         // agy's producer interleaves non-JSON fragments and torn tail writes into its
         // brain jsonl (task 0623: 789/89,818 lines, all unrecoverable). Skipping keeps
         // the 99%+ good records importable without degrading the whole source.

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@gobing-ai/ts-llm-jsonl-importer",
-    "version": "0.4.48",
+    "version": "0.4.50",
     "description": "@gobing-ai/ts-llm-jsonl-importer — Generic JSONL importer for LLM agent history files.",
     "keywords": [
         "typescript",
@@ -47,9 +47,9 @@
         "release": "echo 'Manual publish is disabled. Releases go through GitHub Actions via Trusted Publishing — push a tag: git tag @gobing-ai/ts-llm-jsonl-importer-v<version> && git push --tags' && exit 1"
     },
     "dependencies": {
-        "@gobing-ai/ts-db": "^0.4.48",
-        "@gobing-ai/ts-runtime": "^0.4.48",
-        "@gobing-ai/ts-utils": "^0.4.48",
+        "@gobing-ai/ts-db": "^0.4.50",
+        "@gobing-ai/ts-runtime": "^0.4.50",
+        "@gobing-ai/ts-utils": "^0.4.50",
         "zod": "^4.1.0"
     },
     "devDependencies": {

package/src/mappers.ts

@@ -101,35 +101,129 @@ function redactArgs(args: unknown): RedactedValue {
 }
 
 /**
- * Per-source todo-writing tool names whose raw arguments are retained for phase detection.
- *
- * Evidence: 0489 R4 confirmed omp/pi/claude; codex/grok/agy probed task 0553 R3 from real
- * session JSONL. agy has no on-disk session format (VS Code fork); gemini deferred by the
- * 2026-08-06 source-support ruling. Task 0578 R3: omp `todo_write`, pi `manage_todo_list`,
- * opencode `todowrite`/`todoread` measured live in .spur/spur.db (3,237/21, 1,043, 379/2
- * rows with NULL args_raw).
+ * Retained tool-call arguments: recognized shell tools keep the extracted plain
+ * command (attribution compatibility, capped); every other typed tool call keeps
+ * sanitized-but-valid JSON. `args_digest` is computed elsewhere from the unpruned
+ * args and is unaffected; secret redaction stays at the persistence `redactRecord`
+ * seam (task 0064 R5).
  */
-const TODO_TOOL_ALLOWLIST: Readonly<Record<string, readonly string[]>> = {
-    claude: ['TodoWrite'],
-    opencode: ['todowrite', 'todoread'],
-    codex: ['update_plan'],
-    omp: ['TodoWrite', 'todo', 'todo_write'],
-    grok: ['todo_write'],
-    pi: ['todo', 'manage_todo_list'],
-    agy: [],
-    gemini: [],
-};
+const BASH_COMMAND_MAX_CHARS = 8192;
+const GENERIC_ARGS_MAX_CHARS = 8192;
+
+/** Extract the shell command from a bash tool's raw input; undefined when absent. */
+function bashCommandOf(args: unknown): string | undefined {
+    if (args === null || typeof args !== 'object') return undefined;
+    const rec = args as Record<string, unknown>;
+    const command = rec.command ?? rec.CommandLine ?? rec.cmd ?? rec.script;
+    if (typeof command === 'string') return command;
+    if (Array.isArray(command)) return command.map(String).join(' ');
+    return undefined;
+}
+
+/** Bulky payload keys (normalized) pruned to an omission marker; camelCase and snake_case share one rule. */
+const BULKY_PAYLOAD_KEYS: ReadonlySet<string> = new Set([
+    'content',
+    'codecontent',
+    'replacementcontent',
+    'targetcontent',
+    'filedata',
+    'buffer',
+    'blob',
+    'image',
+    'imagebase64',
+]);
+
+/** Lowercase a key and drop separators so code_content and codeContent compare equal. */
+function normalizeKey(key: string): string {
+    return key.toLowerCase().replace(/[^a-z0-9]/g, '');
+}
+
+/** Tool names whose command-shaped args retain the existing plain-string representation. */
+function isShellTool(toolName: string): boolean {
+    return /^(?:bash|shell|cmd|command|exec|execcommand|executecommand|runcommand|shellcommand|runshellcommand)$/.test(
+        normalizeKey(toolName),
+    );
+}
+
+/** Sanitized tool args: JSON shape preserved, bulky payload strings elided. */
+type SanitizedValue = string | number | boolean | null | SanitizedValue[] | { [key: string]: SanitizedValue };
 
 /**
- * Return JSON-stringified raw args when the tool is a todo-writing tool for its source,
- * otherwise undefined. Codex `arguments` arrives as a JSON string already — store as-is.
+ * Sanitize any JSON value recursively through objects AND arrays: bulky payload
+ * strings under known keys become `[omitted N chars]`, other strings longer than
+ * 1,000 characters keep a prefix plus an explicit original-length marker, and
+ * every primitive and short payload is preserved untouched (task 0064 R2).
  */
-export function maybeArgsRaw(source: string, toolName: string, args: unknown): string | undefined {
-    const allow = TODO_TOOL_ALLOWLIST[source];
-    if (allow === undefined || !allow.includes(toolName)) return undefined;
-    if (typeof args === 'string') return args;
+function sanitizeValue(value: unknown): SanitizedValue {
+    if (typeof value === 'string') {
+        return value.length > 1000 ? `${value.slice(0, 1000)}[truncated ${value.length} chars]` : value;
+    }
+    if (Array.isArray(value)) return value.map(sanitizeValue);
+    if (value !== null && typeof value === 'object') {
+        const out: Record<string, SanitizedValue> = {};
+        for (const [key, entry] of Object.entries(value as Record<string, unknown>)) {
+            out[key] =
+                typeof entry === 'string' && entry.length > 120 && BULKY_PAYLOAD_KEYS.has(normalizeKey(key))
+                    ? `[omitted ${entry.length} chars]`
+                    : sanitizeValue(entry);
+        }
+        return out;
+    }
+    // JSON leaves (number/boolean/null) pass through unchanged; JSON args never
+    // reach here as undefined/bigint/symbol/function.
+    return value as SanitizedValue;
+}
+
+/**
+ * Serialize sanitized args within the retention ceiling without ever slicing JSON
+ * text: values that still serialize past the ceiling collapse into a small valid
+ * JSON wrapper carrying `_truncated`, the omitted character count, and a bounded
+ * preview reduced until the wrapper itself fits (task 0064 R3/AC4).
+ */
+function serializeBounded(value: unknown): string {
+    const serialized = JSON.stringify(sanitizeValue(value));
+    if (serialized.length <= GENERIC_ARGS_MAX_CHARS) return serialized;
+    let preview = serialized;
+    for (;;) {
+        const wrapper = JSON.stringify({
+            _truncated: true,
+            omittedChars: serialized.length - preview.length,
+            preview,
+        });
+        if (wrapper.length <= GENERIC_ARGS_MAX_CHARS) return wrapper;
+        preview = preview.slice(0, preview.length - (wrapper.length - GENERIC_ARGS_MAX_CHARS));
+    }
+}
+
+/**
+ * Return retained raw args for a typed tool call: the shell command string for
+ * recognized bash/shell/exec/command tools, sanitized valid JSON otherwise, and
+ * undefined when the event carries no invocation arguments.
+ */
+export function maybeArgsRaw(_source: string, toolName: string, args: unknown): string | undefined {
     if (args === undefined || args === null) return undefined;
-    return JSON.stringify(args);
+
+    let value = args;
+    if (typeof args === 'string') {
+        // Pre-stringified arguments (Codex) re-enter the structured path so the
+        // retained value parses and prunes like native JSON.
+        try {
+            value = JSON.parse(args);
+        } catch {
+            return args.length > BASH_COMMAND_MAX_CHARS ? args.slice(0, BASH_COMMAND_MAX_CHARS) : args;
+        }
+    }
+
+    if (value !== null && typeof value === 'object') {
+        const cmd = bashCommandOf(value);
+        if (cmd !== undefined && isShellTool(toolName)) {
+            return cmd.length > BASH_COMMAND_MAX_CHARS ? cmd.slice(0, BASH_COMMAND_MAX_CHARS) : cmd;
+        }
+        return serializeBounded(value);
+    }
+
+    // JSON primitives (number/boolean) retain as valid JSON scalars.
+    return JSON.stringify(value);
 }
 
 // ---------------------------------------------------------------------------
@@ -825,68 +919,84 @@ export function codexSplit(raw: Record<string, unknown>, context?: TransformCont
 export function agySplit(raw: Record<string, unknown>, context?: TransformContext): readonly SplitEntry[] {
     const entries: SplitEntry[] = [];
     const recordType = String(raw.type ?? '');
-    const explicitSessionId = s(raw.session_id, raw.conversation_id);
+    const explicitSessionId = s(raw.session_id, raw.conversation_id, raw.conversationId);
     const pathSessionId = context?.sourceFile ? sessionIdFromSourcePath('agy', context.sourceFile) : undefined;
     const sessionId = explicitSessionId ?? pathSessionId ?? 'unknown';
-    const seq = typeof raw.seq === 'number' ? raw.seq : typeof raw.step_index === 'number' ? raw.step_index : 0;
+    const seq =
+        typeof raw.seq === 'number'
+            ? raw.seq
+            : typeof raw.step_index === 'number'
+              ? raw.step_index
+              : (context?.sourceLine ?? 0);
     const ts = timestampOf(raw.created_at, raw.timestamp, raw.ts);
 
     let role: string;
     let disposition: string;
     let contentText: string | null = null;
+    let messageRecordType = recordType;
 
-    // Classification from task 0463 field map (agy transcript types).
-    switch (recordType) {
-        case 'USER_INPUT':
-        case 'ASK_QUESTION':
-            role = 'user';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.text, raw.message) ?? null;
-            break;
-        case 'PLANNER_RESPONSE':
-            role = 'assistant';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.text, raw.message, raw.plan) ?? null;
-            break;
-        case 'ERROR_MESSAGE':
-            role = 'system';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.text, raw.message) ?? null;
-            break;
-        case 'RUN_COMMAND':
-        case 'VIEW_FILE':
-        case 'GREP_SEARCH':
-        case 'LIST_DIRECTORY':
-        case 'READ_URL_CONTENT':
-            role = 'tool';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.command, raw.file_path, raw.pattern) ?? null;
-            break;
-        case 'CODE_ACTION':
-        case 'INVOKE_SUBAGENT':
-            role = 'tool';
-            disposition = 'keep';
-            contentText = s(raw.content, raw.text) ?? null;
-            break;
-        case 'CONVERSATION_HISTORY':
-        case 'CHECKPOINT':
-        case 'GENERIC':
-            role = 'meta';
-            disposition = 'meta';
-            contentText = s(raw.content, raw.text) ?? null;
-            break;
-        default:
-            if (recordType.length === 0) {
-                role = 'unknown';
-                disposition = 'unknown';
-            } else {
-                // Known discriminator present but not in the frozen 0463 keep/meta table —
-                // still a determined type; treat as meta bookkeeping rather than unknown.
+    // history.jsonl prompt index: {display, timestamp, workspace, conversationId?, type?}.
+    // One branch covers absent type, 'slash_command', 'shell', and any future producer type.
+    // `display` never appears on legacy brain transcripts, so this cannot collide with 0463.
+    if (raw.display !== undefined) {
+        role = 'user';
+        disposition = 'keep';
+        contentText = s(raw.display) ?? null;
+        messageRecordType = recordType.length > 0 ? recordType : 'USER_INPUT';
+    } else {
+        // Classification from task 0463 field map (agy transcript types).
+        switch (recordType) {
+            case 'USER_INPUT':
+            case 'ASK_QUESTION':
+                role = 'user';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.text, raw.message) ?? null;
+                break;
+            case 'PLANNER_RESPONSE':
+                role = 'assistant';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.text, raw.message, raw.plan) ?? null;
+                break;
+            case 'ERROR_MESSAGE':
+                role = 'system';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.text, raw.message) ?? null;
+                break;
+            case 'RUN_COMMAND':
+            case 'VIEW_FILE':
+            case 'GREP_SEARCH':
+            case 'LIST_DIRECTORY':
+            case 'READ_URL_CONTENT':
+                role = 'tool';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.command, raw.file_path, raw.pattern) ?? null;
+                break;
+            case 'CODE_ACTION':
+            case 'INVOKE_SUBAGENT':
+                role = 'tool';
+                disposition = 'keep';
+                contentText = s(raw.content, raw.text) ?? null;
+                break;
+            case 'CONVERSATION_HISTORY':
+            case 'CHECKPOINT':
+            case 'GENERIC':
                 role = 'meta';
                 disposition = 'meta';
                 contentText = s(raw.content, raw.text) ?? null;
-            }
-            break;
+                break;
+            default:
+                if (recordType.length === 0) {
+                    role = 'unknown';
+                    disposition = 'unknown';
+                } else {
+                    // Known discriminator present but not in the frozen 0463 keep/meta table —
+                    // still a determined type; treat as meta bookkeeping rather than unknown.
+                    role = 'meta';
+                    disposition = 'meta';
+                    contentText = s(raw.content, raw.text) ?? null;
+                }
+                break;
+        }
     }
 
     const messageSplitIndex = entries.length;
@@ -897,7 +1007,7 @@ export function agySplit(raw: Record<string, unknown>, context?: TransformContex
             session_id: sessionId,
             seq,
             role,
-            record_type: recordType,
+            record_type: messageRecordType,
             disposition,
             ts,
             duration_ms: undefined,
@@ -908,7 +1018,7 @@ export function agySplit(raw: Record<string, unknown>, context?: TransformContex
             cache_write_tokens: null,
             cost_usd: null,
             content_text: contentText,
-            cwd: null,
+            cwd: s(raw.workspace, raw.cwd) ?? null,
             provenance: 'ambient',
         },
     });

package/src/sources.ts

@@ -194,7 +194,7 @@ export const SOURCE_DEFINITIONS: Readonly<Record<LlmJsonlSource, SourceDefinitio
         ...customSourceDefinition(
             'agy',
             'Antigravity CLI',
-            ['.gemini/antigravity-cli/brain'],
+            ['.gemini/antigravity-cli'],
             ['*.jsonl'],
             agySplit,
             AGY_FIELD_MAP,