@akagilnc/pi-workflow-roles 0.1.1941 → 0.1.2004

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@akagilnc/pi-workflow-roles",
-  "version": "0.1.1941",
+  "version": "0.1.2004",
   "description": "Soul-bound workflow roles for Pi",
   "type": "module",
   "license": "Apache-2.0",

package/src/atomic-write.ts

@@ -0,0 +1,26 @@
+/**
+ * Same-directory temp + rename atomic file replace.
+ * Shared primitive for ledger-adjacent typed pages (taishi metrics, etc.).
+ * Does not open/truncate an existing destination inode, so hard-linked twins
+ * keep prior bytes until the directory entry is swapped.
+ * Parent directory must already exist — callers that write under the package
+ * ledger home own confinement via ensureRealDirectoryTree (ADR 0038).
+ */
+import { randomUUID } from "node:crypto";
+import { rename, rm, writeFile } from "node:fs/promises";
+import { dirname, join } from "node:path";
+
+export async function writeFileAtomically(
+  destination: string,
+  contents: string | Uint8Array,
+): Promise<void> {
+  const parent = dirname(destination);
+  const temporary = join(parent, `.atomic-write-${randomUUID()}.tmp`);
+  try {
+    await writeFile(temporary, contents);
+    await rename(temporary, destination);
+  } catch (error) {
+    await rm(temporary, { force: true }).catch(() => undefined);
+    throw error;
+  }
+}

package/src/ledger-session-read.ts

@@ -0,0 +1,260 @@
+/**
+ * Canonical ledger session JSONL read primitives (shared owner).
+ * Consumers (ticket-trajectory, taishi, …) must import here — no second parse kernel.
+ */
+import { readFile } from "node:fs/promises";
+
+export type LedgerSessionRow = Record<string, unknown>;
+
+/**
+ * Loud JSONL failure that still retains rows parsed before the bad line.
+ * Callers that only need the throw keep catching Error; owners that must
+ * surface partial typed facts (e.g. first-frame timestamp) read prefixRows.
+ */
+export class LedgerSessionJsonlError extends Error {
+  readonly path: string;
+  readonly line: number;
+  readonly prefixRows: readonly LedgerSessionRow[];
+
+  constructor(
+    message: string,
+    init: {
+      readonly path: string;
+      readonly line: number;
+      readonly prefixRows: readonly LedgerSessionRow[];
+    },
+  ) {
+    super(message);
+    this.name = "LedgerSessionJsonlError";
+    this.path = init.path;
+    this.line = init.line;
+    this.prefixRows = init.prefixRows;
+  }
+}
+
+function isRecord(value: unknown): value is Record<string, unknown> {
+  return typeof value === "object" && value !== null && !Array.isArray(value);
+}
+
+/**
+ * Read session JSONL with honest live-tail semantics:
+ * a malformed line is tolerated only when it is an unfinished final
+ * fragment at EOF (no record terminator after it). Any malformed line
+ * completed by a line terminator must fail loudly with file and 1-based
+ * line context — even when no non-empty record follows — never silently
+ * under-count.
+ *
+ * Loud failures throw LedgerSessionJsonlError carrying prefixRows so the
+ * single parse kernel can still expose facts obtained before the bad line.
+ */
+export async function readLedgerSessionJsonl(path: string): Promise<LedgerSessionRow[]> {
+  const text = await readFile(path, "utf8");
+  // split keeps a trailing empty segment iff text ends with "\n", so
+  // index < lines.length - 1 means this segment was terminated.
+  const lines = text.split("\n");
+  const rows: LedgerSessionRow[] = [];
+  for (let index = 0; index < lines.length; index += 1) {
+    const line = lines[index]!;
+    if (!line.trim()) continue;
+    let row: unknown;
+    try {
+      row = JSON.parse(line);
+    } catch (error) {
+      if (!(error instanceof SyntaxError)) throw error;
+      const completedByTerminator = index < lines.length - 1;
+      if (completedByTerminator) {
+        throw new LedgerSessionJsonlError(
+          `malformed JSONL record in ${path} at line ${index + 1}: ${error.message}`,
+          { path, line: index + 1, prefixRows: rows },
+        );
+      }
+      // unfinished fragment at EOF — keep prior complete rows
+      break;
+    }
+    // Syntactically complete line: must be a session object. Silent omission
+    // would under-count ledger evidence (failure honesty).
+    if (!isRecord(row)) {
+      const kind = row === null ? "null" : Array.isArray(row) ? "array" : typeof row;
+      throw new LedgerSessionJsonlError(
+        `complete non-object JSONL record in ${path} at line ${index + 1}: expected object, got ${kind}`,
+        { path, line: index + 1, prefixRows: rows },
+      );
+    }
+    rows.push(row);
+  }
+  return rows;
+}
+
+/** First and last record timestamps in encounter order. */
+export function extractSessionTimestampSpan(
+  rows: readonly LedgerSessionRow[],
+): { startedAt?: string; endedAt?: string } {
+  let startedAt: string | undefined;
+  let endedAt: string | undefined;
+  for (const row of rows) {
+    if (typeof row.timestamp !== "string" || !row.timestamp) continue;
+    if (startedAt === undefined) startedAt = row.timestamp;
+    endedAt = row.timestamp;
+  }
+  return {
+    ...(startedAt !== undefined ? { startedAt } : {}),
+    ...(endedAt !== undefined ? { endedAt } : {}),
+  };
+}
+
+/**
+ * Ordered-unique model ids from session frames (first-seen order).
+ * Sources (same faces ticket-trajectory already reads — single parse kernel):
+ * - `model_change.modelId`
+ * - assistant `message.model`
+ * Blank / non-string values are skipped. Does not invent a default model.
+ */
+export function extractSessionModelSequence(
+  rows: readonly LedgerSessionRow[],
+): string[] {
+  const seen = new Set<string>();
+  const ordered: string[] = [];
+  const push = (raw: string): void => {
+    const model = raw.trim();
+    if (model === "" || seen.has(model)) return;
+    seen.add(model);
+    ordered.push(model);
+  };
+  for (const row of rows) {
+    if (row.type === "model_change" && typeof row.modelId === "string") {
+      push(row.modelId);
+    }
+    const message = isRecord(row.message) ? row.message : undefined;
+    if (message?.role === "assistant" && typeof message.model === "string") {
+      push(message.model);
+    }
+  }
+  return ordered;
+}
+
+/** First line of a bash `command` argument (sole owner of this summary). */
+export function bashCommandFirstLine(command: string): string {
+  const match = /^[^\r\n]*/.exec(command);
+  return match?.[0] ?? "";
+}
+
+export type SessionToolInterval = {
+  readonly toolCallId: string;
+  readonly toolName: string;
+  readonly startedAt: string;
+  readonly endedAt?: string;
+  /**
+   * Bash-only first-line command summary from `arguments.command`.
+   * Omitted for non-bash tools and when the argument is absent/non-string.
+   * Full multi-line bodies are never retained on this typed fact face.
+   */
+  readonly command?: string;
+};
+
+/**
+ * Pair toolCall frames → toolResult frames by toolCallId.
+ * Throws when a tool-bearing frame is structurally unreadable for association
+ * (toolCall missing string id, toolResult missing string toolCallId).
+ * Unpaired open calls remain without endedAt — that is incomplete, not unreadable.
+ */
+export function extractSessionToolIntervals(
+  rows: readonly LedgerSessionRow[],
+): SessionToolInterval[] {
+  type Open = {
+    toolCallId: string;
+    toolName: string;
+    startedAt: string;
+    endedAt?: string;
+    command?: string;
+  };
+  const order: Open[] = [];
+  const openById = new Map<string, Open>();
+
+  for (const row of rows) {
+    const rowTimestamp = typeof row.timestamp === "string" ? row.timestamp : undefined;
+    const message = isRecord(row.message) ? row.message : undefined;
+
+    if (message?.role === "assistant" && Array.isArray(message.content)) {
+      const callTimestamp =
+        typeof message.timestamp === "string" && message.timestamp
+          ? message.timestamp
+          : rowTimestamp;
+      for (const part of message.content) {
+        if (!isRecord(part) || part.type !== "toolCall") continue;
+        if (typeof part.id !== "string" || part.id.length === 0) {
+          throw new Error("toolCall frame missing string id");
+        }
+        if (typeof part.name !== "string" || part.name.length === 0) {
+          throw new Error(`toolCall ${part.id} missing string name`);
+        }
+        if (callTimestamp === undefined || callTimestamp.length === 0) {
+          throw new Error(`toolCall ${part.id} missing timestamp`);
+        }
+        if (openById.has(part.id)) {
+          throw new Error(`duplicate toolCall id ${part.id}`);
+        }
+        const args = isRecord(part.arguments) ? part.arguments : undefined;
+        // Ticket surface: only bash first-line summary is authorized here.
+        const command =
+          part.name === "bash" &&
+          args !== undefined &&
+          typeof args.command === "string"
+            ? bashCommandFirstLine(args.command)
+            : undefined;
+        const interval: Open = {
+          toolCallId: part.id,
+          toolName: part.name,
+          startedAt: callTimestamp,
+          ...(command !== undefined ? { command } : {}),
+        };
+        order.push(interval);
+        openById.set(part.id, interval);
+      }
+    }
+
+    if (message?.role === "toolResult") {
+      if (typeof message.toolCallId !== "string" || message.toolCallId.length === 0) {
+        throw new Error("toolResult frame missing string toolCallId");
+      }
+      const resultTimestamp =
+        typeof message.timestamp === "string" && message.timestamp
+          ? message.timestamp
+          : rowTimestamp;
+      if (resultTimestamp === undefined || resultTimestamp.length === 0) {
+        throw new Error(`toolResult ${message.toolCallId} missing timestamp`);
+      }
+      const open = openById.get(message.toolCallId);
+      if (open === undefined) {
+        // Result without a prior call is still associable as a closed interval
+        // once a name is known; keep structural readability without inventing a call.
+        const toolName =
+          typeof message.toolName === "string" && message.toolName.length > 0
+            ? message.toolName
+            : "unknown";
+        order.push({
+          toolCallId: message.toolCallId,
+          toolName,
+          startedAt: resultTimestamp,
+          endedAt: resultTimestamp,
+        });
+        continue;
+      }
+      if (open.endedAt !== undefined) {
+        throw new Error(`duplicate toolResult for toolCallId ${message.toolCallId}`);
+      }
+      open.endedAt = resultTimestamp;
+    }
+  }
+
+  return order.map((interval) => {
+    const base = {
+      toolCallId: interval.toolCallId,
+      toolName: interval.toolName,
+      startedAt: interval.startedAt,
+      ...(interval.command !== undefined ? { command: interval.command } : {}),
+    };
+    return interval.endedAt === undefined
+      ? base
+      : { ...base, endedAt: interval.endedAt };
+  });
+}

package/src/public-cli/cli.ts

@@ -32,6 +32,7 @@ import {
   parseJudgeArgv,
   parseMergerArgv,
   parseReviewerArgv,
+  parseTaishiArgv,
 } from "./invocation.ts";
 import { runPublicCoder, runPublicCoderResume } from "./coder-run.ts";
 import { runPublicCollector } from "./collector-run.ts";
@@ -40,6 +41,7 @@ import { runPublicFixer, runPublicFixerResume } from "./fixer-run.ts";
 import { runPublicJudge, runPublicResume } from "./judge-run.ts";
 import { runPublicMerger, runPublicMergerResume } from "./merger-run.ts";
 import { runPublicReviewer, runPublicReviewerResume } from "./reviewer-run.ts";
+import { runPublicTaishi } from "./taishi-run.ts";
 import { peekRoleRunRole } from "./run-lifecycle.ts";
 import {
   INTERNAL_ROLE_ENTRYPOINT_RELATIVE,
@@ -73,6 +75,8 @@ export const PUBLIC_ROLE_ARGV = {
   doctor: { parse: parseDoctorArgv },
   merger: { parse: parseMergerArgv },
   reviewer: { parse: parseReviewerArgv },
+  /** Deterministic analysis seat (#336) — argv parse only; no LLM admission. */
+  taishi: { parse: parseTaishiArgv },
 } as const;
 
 type TakenPublicGlobalFlag =
@@ -314,6 +318,12 @@ function renderHelp(): string {
       lines.push(`  ${cap.name} — ${phaseText}`);
     }
   }
+  lines.push("", "Deterministic commands:");
+  for (const cap of doc.capabilities) {
+    if (cap.kind === "deterministic") {
+      lines.push(`  ${cap.name}`);
+    }
+  }
   lines.push(
     "",
     "Global options: --model provider/model --thinking level",
@@ -430,6 +440,8 @@ export async function runAkRole(
         }
         if (match.kind === "support") {
           io.stdout(`command\t${match.name}\tkind\tsupport\n`);
+        } else if (match.kind === "deterministic") {
+          io.stdout(`command\t${match.name}\tkind\tdeterministic\n`);
         } else {
           io.stdout(
             `command\t${match.name}\tkind\trole\tphases\t${match.phases
@@ -949,6 +961,18 @@ export async function runAkRole(
       };
     }
 
+    // Taishi public run path: deterministic analysis seat (#336 issue / #337 sweep).
+    // Not an LLM PUBLIC_CALLABLE_ROLE — registered only on PUBLIC_ROLE_ARGV (#176).
+    if (parsed.command === "taishi") {
+      const result = await runPublicTaishi(
+        parsed.args,
+        { home },
+        io,
+        PUBLIC_ROLE_ARGV.taishi.parse,
+      );
+      return { exitCode: result.exitCode };
+    }
+
     // #115: every PUBLIC_CALLABLE_ROLE has a completed handler above. Unknown
     // tokens (including misspelled role names) are structural rejects.
     throw new CliUsageError(`unknown command: ${parsed.command}`);