@gobing-ai/ts-llm-jsonl-importer 0.4.49 → 0.4.50

package/dist/mappers.d.ts

@@ -3,12 +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 retained raw args for tools whose arguments carry forensic value, otherwise
- * undefined. Todo-writing tools retain the full JSON args (phase detection); bash tools
- * retain just the command string (capped). Codex `arguments` arrives as a JSON string
- * already — store as-is for todo tools.
+ * 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;AA6ED;;;;;GAKG;AACH,wBAAgB,YAAY,CAAC,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,EAAE,IAAI,EAAE,OAAO,GAAG,MAAM,GAAG,SAAS,CAchG;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"}
+{"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

@@ -91,79 +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).
- */
-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: [],
-};
-/**
- * Per-source shell/exec tool names whose command string is retained in args_raw.
- *
- * WHY: bash commands are the primary forensic signal for session-task attribution and ops
- * history (spur's task classifier regexes `spur task <verb> <wbs>` out of args_raw). Only
- * evidenced sources are listed; unverified shapes stay empty the same way the todo allowlist
- * treats agy/gemini. Commands pass through the importer's redaction layer like any other
- * retained field (importer.ts redacts every split record before insert).
+ * 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 BASH_TOOL_ALLOWLIST = {
-    claude: ['Bash'], // input: { command: string } (mappers.test.ts fixture)
-    opencode: ['bash'], // input: { command: string } (opencode-importer.test.ts)
-    codex: [], // arguments arrive as a pre-stringified JSON string; command shape unverified
-    omp: [],
-    grok: [],
-    pi: ['bash'], // arguments: { command: string } (live pi session transcripts)
-    agy: [],
-    gemini: [],
-};
-/** Retained bash commands are capped; attribution only needs the invocation line. */
 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 command = args.command;
+    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)$/.test(normalizeKey(toolName));
+}
 /**
- * Return retained raw args for tools whose arguments carry forensic value, otherwise
- * undefined. Todo-writing tools retain the full JSON args (phase detection); bash tools
- * retain just the command string (capped). Codex `arguments` arrives as a JSON string
- * already — store as-is for todo tools.
+ * Sanitize any JSON value recursively through objects AND arrays: bulky payload
+ * strings under known keys become `[omitted N chars]`, other strings longer than
+ * 1,000 characters keep a prefix plus an explicit original-length marker, and
+ * every primitive and short payload is preserved untouched (task 0064 R2).
  */
-export function maybeArgsRaw(source, toolName, args) {
-    const todo = TODO_TOOL_ALLOWLIST[source];
-    if (todo?.includes(toolName)) {
-        if (typeof args === 'string')
-            return args;
-        if (args === undefined || args === null)
-            return undefined;
-        return JSON.stringify(args);
+function sanitizeValue(value) {
+    if (typeof value === 'string') {
+        return value.length > 1000 ? `${value.slice(0, 1000)}[truncated ${value.length} chars]` : value;
     }
-    const bash = BASH_TOOL_ALLOWLIST[source];
-    if (bash?.includes(toolName)) {
-        const command = bashCommandOf(args);
-        if (command === undefined)
-            return undefined;
-        return command.length > BASH_COMMAND_MAX_CHARS ? command.slice(0, BASH_COMMAND_MAX_CHARS) : command;
+    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;
     }
-    return undefined;
+    // 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;
+    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

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@gobing-ai/ts-llm-jsonl-importer",
-    "version": "0.4.49",
+    "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": "^0.4.49",
-        "@gobing-ai/ts-runtime": "^0.4.49",
-        "@gobing-ai/ts-utils": "^0.4.49",
+        "@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

@@ -101,77 +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).
- */
-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: [],
-};
-
-/**
- * Per-source shell/exec tool names whose command string is retained in args_raw.
- *
- * WHY: bash commands are the primary forensic signal for session-task attribution and ops
- * history (spur's task classifier regexes `spur task <verb> <wbs>` out of args_raw). Only
- * evidenced sources are listed; unverified shapes stay empty the same way the todo allowlist
- * treats agy/gemini. Commands pass through the importer's redaction layer like any other
- * retained field (importer.ts redacts every split record before insert).
+ * 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 BASH_TOOL_ALLOWLIST: Readonly<Record<string, readonly string[]>> = {
-    claude: ['Bash'], // input: { command: string } (mappers.test.ts fixture)
-    opencode: ['bash'], // input: { command: string } (opencode-importer.test.ts)
-    codex: [], // arguments arrive as a pre-stringified JSON string; command shape unverified
-    omp: [],
-    grok: [],
-    pi: ['bash'], // arguments: { command: string } (live pi session transcripts)
-    agy: [],
-    gemini: [],
-};
-
-/** Retained bash commands are capped; attribution only needs the invocation line. */
 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 command = (args as Record<string, unknown>).command;
+    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)$/.test(
+        normalizeKey(toolName),
+    );
+}
+
+/** Sanitized tool args: JSON shape preserved, bulky payload strings elided. */
+type SanitizedValue = string | number | boolean | null | SanitizedValue[] | { [key: string]: SanitizedValue };
+
 /**
- * Return retained raw args for tools whose arguments carry forensic value, otherwise
- * undefined. Todo-writing tools retain the full JSON args (phase detection); bash tools
- * retain just the command string (capped). Codex `arguments` arrives as a JSON string
- * already — store as-is for todo tools.
+ * Sanitize any JSON value recursively through objects AND arrays: bulky payload
+ * strings under known keys become `[omitted N chars]`, other strings longer than
+ * 1,000 characters keep a prefix plus an explicit original-length marker, and
+ * every primitive and short payload is preserved untouched (task 0064 R2).
  */
-export function maybeArgsRaw(source: string, toolName: string, args: unknown): string | undefined {
-    const todo = TODO_TOOL_ALLOWLIST[source];
-    if (todo?.includes(toolName)) {
-        if (typeof args === 'string') return args;
-        if (args === undefined || args === null) return undefined;
-        return JSON.stringify(args);
+function sanitizeValue(value: unknown): SanitizedValue {
+    if (typeof value === 'string') {
+        return value.length > 1000 ? `${value.slice(0, 1000)}[truncated ${value.length} chars]` : value;
     }
-    const bash = BASH_TOOL_ALLOWLIST[source];
-    if (bash?.includes(toolName)) {
-        const command = bashCommandOf(args);
-        if (command === undefined) return undefined;
-        return command.length > BASH_COMMAND_MAX_CHARS ? command.slice(0, BASH_COMMAND_MAX_CHARS) : command;
+    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;
     }
-    return undefined;
+    // 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;
+
+    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);
 }
 
 // ---------------------------------------------------------------------------