@runtypelabs/cli 2.1.0 → 2.2.0

package/dist/index.js

@@ -8469,8 +8469,12 @@ function setStoredToolPayloadField(payloads, toolId, field, value) {
   const next = { ...payloads.get(toolId) ?? {} };
   if (value === void 0) {
     delete next[field];
+  } else if (field === "parameters") {
+    next.parameters = value;
+  } else if (field === "result") {
+    next.result = value;
   } else {
-    next[field] = value;
+    next.streamedInput = value;
   }
   if (next.parameters === void 0 && next.result === void 0 && next.streamedInput === void 0) {
     payloads.delete(toolId);
@@ -14797,6 +14801,8 @@ var NETWORK_ERROR_PATTERNS = [
   "econnrefused",
   "econnaborted",
   "etimedout",
+  "timeout",
+  "request timeout",
   "enetunreach",
   "enetdown",
   "ehostunreach",
@@ -14816,12 +14822,78 @@ var NETWORK_ERROR_PATTERNS = [
   "unable to connect",
   "err_network"
 ];
+var LOCAL_NETWORK_PATTERNS = [
+  "enetunreach",
+  "enetdown",
+  "enotfound",
+  "network error",
+  "network request failed",
+  "networkerror",
+  "err_network"
+];
+var SERVER_UNREACHABLE_PATTERNS = [
+  "econnrefused",
+  "econnreset",
+  "connection refused",
+  "connection reset",
+  "ehostunreach"
+];
+function collectErrorSignals(error, seen = /* @__PURE__ */ new Set()) {
+  if (error == null || seen.has(error)) return [];
+  if (typeof error === "string") return [error];
+  if (typeof error !== "object") return [String(error)];
+  seen.add(error);
+  const parts = [];
+  if ("message" in error && typeof error.message === "string") {
+    parts.push(error.message);
+  }
+  if ("code" in error && typeof error.code === "string") {
+    parts.push(error.code);
+  }
+  if ("cause" in error) {
+    parts.push(...collectErrorSignals(error.cause, seen));
+  }
+  return parts;
+}
+function getNetworkErrorContext(error) {
+  const signals = collectErrorSignals(error);
+  const fallbackMessage = error instanceof Error ? error.message : String(error);
+  const uniqueSignals = [...new Set(signals.map((signal) => signal.trim()).filter(Boolean))];
+  const searchText = uniqueSignals.join(" ").toLowerCase();
+  const detailMessage = uniqueSignals.find((signal) => signal.toLowerCase() !== "fetch failed") ?? fallbackMessage;
+  return {
+    searchText,
+    detailMessage
+  };
+}
+function describeNetworkError(error) {
+  const { searchText, detailMessage } = getNetworkErrorContext(error);
+  const isLocalNetwork = LOCAL_NETWORK_PATTERNS.some((p) => searchText.includes(p));
+  const isServerUnreachable = SERVER_UNREACHABLE_PATTERNS.some((p) => searchText.includes(p));
+  const isTimeout = searchText.includes("etimedout") || searchText.includes("timeout");
+  const lines = [];
+  if (isLocalNetwork) {
+    lines.push("Could not reach the Runtype API \u2014 your network appears to be offline.");
+    lines.push("Check your internet connection and try again.");
+  } else if (isServerUnreachable) {
+    lines.push("Could not reach the Runtype API \u2014 the server is not responding.");
+    lines.push("The service may be temporarily unavailable. Try again in a few minutes.");
+  } else if (isTimeout) {
+    lines.push("Could not reach the Runtype API \u2014 the request timed out.");
+    lines.push("This could be a network issue or the server may be under heavy load.");
+  } else {
+    lines.push("Could not reach the Runtype API \u2014 a network error occurred.");
+    lines.push("Check your internet connection or try again in a few minutes.");
+  }
+  lines.push(`Details: ${detailMessage}`);
+  return lines;
+}
 function isTransientNetworkError(error) {
   if (error instanceof RuntypeApiError) return false;
-  const message = error instanceof Error ? error.message.toLowerCase() : String(error).toLowerCase();
-  if (error instanceof TypeError && message.includes("fetch")) return true;
+  const { searchText } = getNetworkErrorContext(error);
+  if (error instanceof TypeError && searchText.includes("fetch")) return true;
   if (error instanceof DOMException && error.name === "AbortError") return true;
-  return NETWORK_ERROR_PATTERNS.some((pattern) => message.includes(pattern));
+  return NETWORK_ERROR_PATTERNS.some((pattern) => searchText.includes(pattern));
 }
 async function retryOnNetworkError(fn, opts = {}) {
   const maxRetries = opts.maxRetries ?? 3;
@@ -14898,9 +14970,8 @@ function describeMarathonApiError(error) {
   if (!(error instanceof Error)) {
     return ["Task failed: Unknown error"];
   }
-  if (!(error instanceof RuntypeApiError) || error.statusCode !== 429) {
-    const message = error instanceof Error ? error.message : "Unknown error";
-    return [`Task failed: ${message}`];
+  if (isTransientNetworkError(error)) {
+    return describeNetworkError(error);
   }
   return [`Task failed: ${error.message}`];
 }
@@ -16134,6 +16205,57 @@ function createRunCheckTool() {
     }
   };
 }
+function createSearchSessionHistoryTool(client, taskName) {
+  return {
+    description: "Search across all prior marathon sessions for specific information, decisions, findings, or tool outputs. Use this when you need to recall something from earlier sessions that may have been compacted away. Returns ranked results with content snippets from matching sessions.",
+    parametersSchema: {
+      type: "object",
+      properties: {
+        query: {
+          type: "string",
+          description: 'What to search for (e.g. "authentication flow decisions", "test failures in auth module")'
+        },
+        limit: {
+          type: "number",
+          description: "Maximum number of results to return (default 5, max 20)"
+        },
+        types: {
+          type: "array",
+          items: { type: "string", enum: ["response", "reasoning", "tool_output"] },
+          description: "Filter by content type (default: all types)"
+        }
+      },
+      required: ["query"]
+    },
+    execute: async (args) => {
+      const query = String(args.query || "").trim();
+      if (!query) return "Error: query is required";
+      const limit = Math.max(1, Math.min(20, Number(args.limit) || 5));
+      const types = Array.isArray(args.types) ? args.types : void 0;
+      try {
+        const response = await client.post("/session-context/search", {
+          query,
+          taskName,
+          limit,
+          ...types ? { types } : {}
+        });
+        if (!response.success || !response.results || response.results.length === 0) {
+          return "No matching session context found for your query.";
+        }
+        const formatted = response.results.map((r, i) => {
+          const header = `[Result ${i + 1}] Session ${r.sessionIndex} | ${r.type}${r.toolName ? ` (${r.toolName})` : ""} | Score: ${r.score.toFixed(3)}`;
+          return `${header}
+${r.content}`;
+        });
+        return `Found ${response.count} matching results:
+
+${formatted.join("\n\n---\n\n")}`;
+      } catch (error) {
+        return `Session search unavailable: ${error instanceof Error ? error.message : String(error)}`;
+      }
+    }
+  };
+}
 function buildLocalTools(client, sandboxProvider, options, context) {
   const enabledTools = {};
   if (!options.noLocalTools) {
@@ -16149,6 +16271,9 @@ function buildLocalTools(client, sandboxProvider, options, context) {
         context.stateDir
       );
       enabledTools.run_check = createRunCheckTool();
+      if (options.sessionSearch === true) {
+        enabledTools.search_session_history = createSearchSessionHistoryTool(client, context.taskName);
+      }
     }
   }
   if (sandboxProvider) {
@@ -16160,6 +16285,62 @@ function buildLocalTools(client, sandboxProvider, options, context) {
   return Object.keys(enabledTools).length > 0 ? enabledTools : void 0;
 }
 
+// src/marathon/session-chunker.ts
+var DEFAULT_MAX_CHUNK_CHARS = 2e3;
+var MIN_CONTENT_LENGTH = 50;
+function extractSessionChunks(snapshot, maxChunkChars = DEFAULT_MAX_CHUNK_CHARS) {
+  const chunks = [];
+  if (snapshot.content && snapshot.content.length >= MIN_CONTENT_LENGTH) {
+    chunks.push(...chunkText(snapshot.content, "response", maxChunkChars));
+  }
+  if (snapshot.reasoning && snapshot.reasoning.length >= MIN_CONTENT_LENGTH) {
+    chunks.push(...chunkText(snapshot.reasoning, "reasoning", maxChunkChars));
+  }
+  for (const tool of snapshot.tools) {
+    const result = typeof tool.result === "string" ? tool.result : JSON.stringify(tool.result ?? "");
+    if (result.length >= MIN_CONTENT_LENGTH) {
+      chunks.push(
+        ...chunkText(result, "tool_output", maxChunkChars, tool.name)
+      );
+    }
+  }
+  return chunks;
+}
+function chunkText(text, type, maxChars, toolName) {
+  if (text.length <= maxChars) {
+    return [{ content: text, type, ...toolName ? { toolName } : {} }];
+  }
+  const chunks = [];
+  const paragraphs = text.split(/\n\n+/);
+  let current = "";
+  for (const paragraph of paragraphs) {
+    if (paragraph.length > maxChars) {
+      if (current.length >= MIN_CONTENT_LENGTH) {
+        chunks.push({ content: current.trim(), type, ...toolName ? { toolName } : {} });
+        current = "";
+      }
+      const sentences = paragraph.match(/[^.!?]+[.!?]+\s*|[^.!?]+$/g) || [paragraph];
+      for (const sentence of sentences) {
+        if (current.length + sentence.length > maxChars && current.length > 0) {
+          chunks.push({ content: current.trim(), type, ...toolName ? { toolName } : {} });
+          current = "";
+        }
+        current += sentence;
+      }
+      continue;
+    }
+    if (current.length + paragraph.length + 2 > maxChars && current.length >= MIN_CONTENT_LENGTH) {
+      chunks.push({ content: current.trim(), type, ...toolName ? { toolName } : {} });
+      current = "";
+    }
+    current += (current ? "\n\n" : "") + paragraph;
+  }
+  if (current.length >= MIN_CONTENT_LENGTH) {
+    chunks.push({ content: current.trim(), type, ...toolName ? { toolName } : {} });
+  }
+  return chunks;
+}
+
 // src/marathon/loop-detector.ts
 var DEFAULT_MAX_HISTORY = 30;
 var DEFAULT_MIN_PATTERN_LENGTH = 2;
@@ -17090,6 +17271,9 @@ async function taskAction(agent, options) {
             console.log(chalk16.green(`Created agent: ${agentId}`));
           }
         } catch (createErr) {
+          if (isTransientNetworkError(createErr)) {
+            await failBeforeMain(formatMarathonApiError(createErr));
+          }
           const errMsg = createErr instanceof Error ? createErr.message : String(createErr);
           await failBeforeMain([
             chalk16.red(`Failed to create agent "${normalizedAgent}"`),
@@ -17098,6 +17282,9 @@ async function taskAction(agent, options) {
         }
       }
     } catch (error) {
+      if (isTransientNetworkError(error)) {
+        await failBeforeMain(formatMarathonApiError(error));
+      }
       const errMsg = error instanceof Error ? error.message : String(error);
       await failBeforeMain([
         chalk16.red("Failed to list agents"),
@@ -17752,6 +17939,21 @@ Saving state... done. Session saved to ${filePath}`);
           resumeState = extractRunTaskResumeState(adjustedState);
           lastSessionMessages = state.messages ?? [];
           saveState(filePath, adjustedState, { stripSnapshotEvents: !!eventLogWriter });
+          if (options.sessionSearch === true) {
+            const latestSnapshot = persistedSessionSnapshots[persistedSessionSnapshots.length - 1];
+            if (latestSnapshot) {
+              const chunks = extractSessionChunks(latestSnapshot);
+              if (chunks.length > 0) {
+                const sessionIdx = currentSessionOffset + state.sessionCount - 1;
+                client.post("/session-context/index", {
+                  taskName,
+                  sessionIndex: sessionIdx,
+                  chunks
+                }).catch(() => {
+                });
+              }
+            }
+          }
           if (resumeState?.workflowPhase) {
             const displayMilestone = detectedVariant === "external" && resumeState.workflowPhase === "research" && adjustedState.planWritten ? "report" : resumeState.workflowPhase;
             streamRef.current?.updateMilestone(displayMilestone);
@@ -18127,7 +18329,7 @@ ${details}`);
   }
   return resolved;
 }
-function detectDeployWorkflow(_message, sandboxProvider, resumeState) {
+function detectDeployWorkflow(_message, _sandboxProvider, resumeState) {
   if (resumeState?.workflowVariant === "game") return gameWorkflow;
   if (resumeState?.workflowPhase === "design" || resumeState?.workflowPhase === "build" || resumeState?.workflowPhase === "verify") {
     return gameWorkflow;
@@ -18158,7 +18360,7 @@ function resolveSandboxWorkflowSelection(message, sandboxProvider, resumeState)
   };
 }
 function applyTaskOptions(cmd) {
-  return cmd.argument("<agent>", "Agent ID or name").option("-g, --goal <text>", "Goal message for the agent").option("--max-sessions <n>", "Maximum sessions", "50").option("--max-cost <n>", "Budget in USD").option("--model <modelId>", "Model ID to use (overrides agent config)").option("--name <name>", "Task name (used for state file, defaults to agent name)").option("--session <name>", "Resume a specific session by name").option("--state-dir <path>", "Directory for state files (default: ~/.runtype/projects/<hash>/marathons/)").option("--resume [message]", "Resume from existing local state, optionally with a new message").option("--fresh", "Start a new run and ignore any existing local state for this task").option("--compact", "Force compact-summary resume mode instead of replaying full history").option("--compact-strategy <strategy>", "Compaction strategy: auto (default), provider_native, or summary_fallback").option("--compact-threshold <value>", "Auto-compact when estimated context crosses this threshold (default: 80% fallback, 90% native; accepts percent like 90% or absolute token count like 120000)").option("--compact-instructions <text>", "Extra instructions for what a compact summary must preserve").option("--no-auto-compact", "Disable automatic context-aware history compaction").option("--track", "Sync progress to a Runtype record (visible in dashboard)").option("--debug", "Show debug output from each session").option("--json", "Output final result as JSON").option("--sandbox <provider>", "Enable sandbox code execution tool (cloudflare-worker, quickjs, or daytona)").option("--no-local-tools", "Disable built-in local tool execution (read_file, write_file, list_directory)").option("-t, --tools <tools...>", "Enable built-in tools (e.g., exa, firecrawl, dalle, openai_web_search, anthropic_web_search)").option("--plain-text", "Disable markdown rendering in output").option("--no-reasoning", "Disable model reasoning/thinking (enabled by default for supported models)").option("--no-checkpoint", "Run all iterations without checkpoint pauses (fully autonomous)").option("--checkpoint-timeout <seconds>", "Auto-continue timeout in seconds (default: 10)", "10").option("--planning-model <modelId>", "Model to use during research/planning phases").option("--execution-model <modelId>", "Model to use during execution phase").option("--fallback-model <modelId>", "Model to fall back to when primary model fails").option("--playbook <name>", "Load a playbook from .runtype/marathons/playbooks/").option("--offload-threshold <chars>", 'Offload tool outputs larger than this to files (default: 100000; use "off" or "0" to disable guardrails)').option("--tool-context <mode>", "Tool result storage: hot-tail (default), observation-mask, or full-inline").option("--tool-window <window>", 'Compaction window: "session" (default) or a number for last-N tool results (e.g. 10)').option("--runner-char <char>", "Custom runner emoji (default: \u{1F3C3})").option("--finish-char <char>", "Custom finish line emoji (default: \u{1F3C1})").option("--no-runner", "Hide the runner emoji from the header border").option("--no-finish", "Hide the finish line emoji from the header border").action(taskAction);
+  return cmd.argument("<agent>", "Agent ID or name").option("-g, --goal <text>", "Goal message for the agent").option("--max-sessions <n>", "Maximum sessions", "50").option("--max-cost <n>", "Budget in USD").option("--model <modelId>", "Model ID to use (overrides agent config)").option("--name <name>", "Task name (used for state file, defaults to agent name)").option("--session <name>", "Resume a specific session by name").option("--state-dir <path>", "Directory for state files (default: ~/.runtype/projects/<hash>/marathons/)").option("--resume [message]", "Resume from existing local state, optionally with a new message").option("--fresh", "Start a new run and ignore any existing local state for this task").option("--compact", "Force compact-summary resume mode instead of replaying full history").option("--compact-strategy <strategy>", "Compaction strategy: auto (default), provider_native, or summary_fallback").option("--compact-threshold <value>", "Auto-compact when estimated context crosses this threshold (default: 80% fallback, 90% native; accepts percent like 90% or absolute token count like 120000)").option("--compact-instructions <text>", "Extra instructions for what a compact summary must preserve").option("--no-auto-compact", "Disable automatic context-aware history compaction").option("--track", "Sync progress to a Runtype record (visible in dashboard)").option("--debug", "Show debug output from each session").option("--json", "Output final result as JSON").option("--sandbox <provider>", "Enable sandbox code execution tool (cloudflare-worker, quickjs, or daytona)").option("--no-local-tools", "Disable built-in local tool execution (read_file, write_file, list_directory)").option("--session-search", "Enable session context indexing and search_session_history tool").option("-t, --tools <tools...>", "Enable built-in tools (e.g., exa, firecrawl, dalle, openai_web_search, anthropic_web_search)").option("--plain-text", "Disable markdown rendering in output").option("--no-reasoning", "Disable model reasoning/thinking (enabled by default for supported models)").option("--no-checkpoint", "Run all iterations without checkpoint pauses (fully autonomous)").option("--checkpoint-timeout <seconds>", "Auto-continue timeout in seconds (default: 10)", "10").option("--planning-model <modelId>", "Model to use during research/planning phases").option("--execution-model <modelId>", "Model to use during execution phase").option("--fallback-model <modelId>", "Model to fall back to when primary model fails").option("--playbook <name>", "Load a playbook from .runtype/marathons/playbooks/").option("--offload-threshold <chars>", 'Offload tool outputs larger than this to files (default: 100000; use "off" or "0" to disable guardrails)').option("--tool-context <mode>", "Tool result storage: hot-tail (default), observation-mask, or full-inline").option("--tool-window <window>", 'Compaction window: "session" (default) or a number for last-N tool results (e.g. 10)').option("--runner-char <char>", "Custom runner emoji (default: \u{1F3C3})").option("--finish-char <char>", "Custom finish line emoji (default: \u{1F3C1})").option("--no-runner", "Hide the runner emoji from the header border").option("--no-finish", "Hide the finish line emoji from the header border").action(taskAction);
 }
 var taskCommand = applyTaskOptions(
   new Command10("task").description("Run a multi-session agent task")