feat: improve tool usage guidance in SOUL.md and add cron.create/cron.delete tools

- SOUL.md: list all available tools (web.search, memory.*, cron.*, etc.)
  and add Tool Usage Rules section enforcing 'act, don't narrate'
- cron.ts: add getJob(), addJob(), removeJob() to CronScheduler for
  runtime (ephemeral) cron job management
- cron tools: add cron.create and cron.delete tools, enhance cron.list
  to show schedule/output/message details
- policy.ts: add cron tools to messaging and coding profiles, add
  group:cron to tool groups

Fixes issue where models would narrate tool intent ('let me search...')
then stop without actually calling tools.

SOUL.md

@@ -48,9 +48,24 @@ You have tools for interacting with your operator's system:
 - **file.patch** -- Apply structured multi-hunk patches to one or more files. Line-based replacements, insertions, and deletions in a single call.
 - **file.list** -- List directory contents. Supports glob patterns.
 - **system.info** -- Get current date, time, hostname, platform, and system information.
-- **web.fetch** -- Fetch web pages. Use for looking things up, checking URLs, downloading content.
+- **web.fetch** -- Fetch a specific URL and extract its content as markdown or text. Use when you already know the URL.
+- **web.search** -- Search the web for current information. Returns titles, URLs, and snippets. Use this to look things up when you don't have a specific URL.
+- **memory.read / memory.write / memory.search** -- Persistent memory across sessions. Store and retrieve notes, facts, and context.
+- **cron.list / cron.trigger / cron.create** -- List, manually trigger, or create new scheduled automation jobs.
+- **process.start / process.status / process.output / process.kill / process.list** -- Manage background processes.
+- **message.send** -- Send messages to other channels (Telegram, Discord, etc.).
 
-Use tools when the task requires it. For conversational questions, respond directly. Don't narrate tool usage -- just use them and present results.
+Additional tools (image.analyze, media.send, browser.*, gmail.*, calendar.*, sessions.*, agents.list) may be available depending on configuration. Check your tool definitions if unsure.
+
+## Tool Usage Rules
+
+**Act, don't narrate.** When a task requires tools, call them immediately. Never say "let me search for that" or "I'll look that up" and then stop -- actually call the tool in the same response. The worst possible behavior is describing what you would do without doing it.
+
+**Be honest about limitations.** If you lack the tools or access to complete a task, say so clearly. Never generate a confident-sounding response that implies you're about to take action when you have no way to follow through. "I don't have a tool for that" is always better than "Let me do that for you" followed by nothing.
+
+**Use the right tool.** If someone asks you to search for something and you have `web.search`, use it. Don't fall back to `web.fetch` with a guessed URL. If someone asks about recent events, search -- don't guess from training data.
+
+For conversational questions, respond directly. Don't narrate tool usage -- just use them and present results.
 
 ---
 

docs/plans/state.json

@@ -1063,7 +1063,7 @@
   },
 
   "overall_progress": {
-    "total_test_count": 1292,
+    "total_test_count": 1329,
     "all_tests_passing": true,
     "p0_completion": "3/3 (100%)",
     "p1_completion": "4/4 (100%)",
@@ -1082,5 +1082,15 @@
     "operator_dx_milestone": "Phase 3 (Live Ops Dashboard): 1/2 plans complete — metrics backend done, dashboard UI next",
     "gmail_auth_cli": "flynn gmail-auth command implemented with OAuth2 flow, doctor check, config routed to Telegram",
     "next_up": "GSD Milestone: Operator DX — Phase 3 Plan 02 (Dashboard UI consuming metrics RPC). All phases P0-P8 and Tiers 1-4 complete. Setup wizard added. TUI fullscreen mode now has full tool access and proper display. Remaining gaps: Tier 4 channels (Signal, Matrix, Teams, Google Chat), Tier 5 deferred/niche items"
+  },
+  "soul_md_and_cron_create": {
+    "date": "2026-02-11",
+    "summary": "SOUL.md updated with comprehensive tool listing (web.search, memory.*, cron.*, process.*) and stronger 'act don't narrate' guidance. Added cron.create and cron.delete tools for runtime cron job management. Updated tool policy profiles (messaging/coding) and group:cron. Fixed issue where Flynn would narrate intended actions without calling tools.",
+    "files_modified": [
+      "SOUL.md",
+      "src/tools/builtin/cron.ts",
+      "src/automation/cron.ts",
+      "src/tools/policy.ts"
+    ]
   }
 }

src/automation/cron.ts

@@ -100,4 +100,54 @@ export class CronScheduler implements ChannelAdapter {
   getJobNames(): string[] {
     return Array.from(this.jobs.keys());
   }
+
+  /** Get a job's config by name. */
+  getJob(name: string): CronJobConfig | undefined {
+    return this.jobs.get(name);
+  }
+
+  /**
+   * Add a new cron job at runtime and start it immediately.
+   * The job is ephemeral — it persists only until the daemon restarts.
+   * Returns true if the job was created, false if a job with that name already exists.
+   */
+  addJob(config: CronJobConfig): boolean {
+    if (this.jobs.has(config.name)) {
+      return false;
+    }
+
+    this.jobs.set(config.name, config);
+
+    if (config.enabled && this._status === 'connected') {
+      const cronInstance = new Cron(config.schedule, {
+        timezone: config.timezone,
+        paused: false,
+      }, () => {
+        this.triggerJob(config.name);
+      });
+
+      this.cronInstances.set(config.name, cronInstance);
+    }
+
+    return true;
+  }
+
+  /**
+   * Remove a cron job by name. Stops the cron instance if running.
+   * Returns true if the job was found and removed, false otherwise.
+   */
+  removeJob(name: string): boolean {
+    if (!this.jobs.has(name)) {
+      return false;
+    }
+
+    const cronInstance = this.cronInstances.get(name);
+    if (cronInstance) {
+      cronInstance.stop();
+      this.cronInstances.delete(name);
+    }
+
+    this.jobs.delete(name);
+    return true;
+  }
 }

src/tools/builtin/cron.ts

@@ -8,7 +8,7 @@ export function createCronTools(scheduler: CronScheduler): Tool[] {
   const cronList: Tool = {
     name: 'cron.list',
     description:
-      'List all configured cron jobs with their names and status.',
+      'List all configured cron jobs with their names, schedules, and messages.',
     inputSchema: {
       type: 'object',
       properties: {},
@@ -21,7 +21,11 @@ export function createCronTools(scheduler: CronScheduler): Tool[] {
           return { success: true, output: 'No cron jobs configured.' };
         }
 
-        const lines = jobNames.map((name) => `- ${name}`);
+        const lines = jobNames.map((name) => {
+          const job = scheduler.getJob(name);
+          if (!job) return `- ${name}`;
+          return `- **${name}** — schedule: \`${job.schedule}\`, enabled: ${job.enabled}, output: ${job.output.channel}/${job.output.peer}\n  message: "${job.message.length > 80 ? job.message.slice(0, 80) + '...' : job.message}"`;
+        });
         return {
           success: true,
           output: `${jobNames.length} cron job(s):\n\n${lines.join('\n')}`,
@@ -78,5 +82,137 @@ export function createCronTools(scheduler: CronScheduler): Tool[] {
     },
   };
 
-  return [cronList, cronTrigger];
+  const cronCreate: Tool = {
+    name: 'cron.create',
+    description:
+      'Create a new cron job that sends a message on a schedule. The job starts immediately and runs until the daemon restarts (ephemeral). Use standard cron syntax for the schedule (e.g. "0 9 * * *" for daily at 9am, "*/30 * * * *" for every 30 minutes).',
+    inputSchema: {
+      type: 'object',
+      properties: {
+        name: {
+          type: 'string',
+          description: 'Unique name for the cron job (kebab-case, e.g. "apartment-check")',
+        },
+        schedule: {
+          type: 'string',
+          description: 'Cron schedule expression (e.g. "0 9 * * *" for daily at 9am, "0 */6 * * *" for every 6 hours)',
+        },
+        message: {
+          type: 'string',
+          description: 'The message to send when the job fires. This message will be processed as if a user sent it, so you can include instructions for yourself.',
+        },
+        output_channel: {
+          type: 'string',
+          description: 'Channel to send output to (e.g. "telegram", "discord", "slack"). Defaults to "cron" (responses stay in the cron session).',
+        },
+        output_peer: {
+          type: 'string',
+          description: 'Peer ID on the output channel (e.g. Telegram chat ID). Required if output_channel is set.',
+        },
+        timezone: {
+          type: 'string',
+          description: 'IANA timezone (e.g. "America/Los_Angeles"). Defaults to system timezone.',
+        },
+      },
+      required: ['name', 'schedule', 'message'],
+    },
+    execute: async (rawArgs: unknown): Promise<ToolResult> => {
+      const args = rawArgs as {
+        name: string;
+        schedule: string;
+        message: string;
+        output_channel?: string;
+        output_peer?: string;
+        timezone?: string;
+      };
+
+      try {
+        // Validate name format
+        if (!/^[a-z0-9][a-z0-9-]*[a-z0-9]$|^[a-z0-9]$/.test(args.name)) {
+          return {
+            success: false,
+            output: '',
+            error: 'Job name must be kebab-case (lowercase letters, numbers, and hyphens, not starting/ending with hyphen)',
+          };
+        }
+
+        const outputChannel = args.output_channel ?? 'cron';
+        const outputPeer = args.output_peer ?? args.name;
+
+        const created = scheduler.addJob({
+          name: args.name,
+          schedule: args.schedule,
+          message: args.message,
+          output: {
+            channel: outputChannel,
+            peer: outputPeer,
+          },
+          enabled: true,
+          timezone: args.timezone,
+        });
+
+        if (!created) {
+          return {
+            success: false,
+            output: '',
+            error: `Cron job "${args.name}" already exists. Use a different name or delete the existing one first.`,
+          };
+        }
+
+        return {
+          success: true,
+          output: `Cron job "${args.name}" created and started.\nSchedule: ${args.schedule}${args.timezone ? ` (${args.timezone})` : ''}\nOutput: ${outputChannel}/${outputPeer}\nMessage: "${args.message}"\n\nNote: This job is ephemeral and will not persist across daemon restarts. To make it permanent, add it to the config file.`,
+        };
+      } catch (error) {
+        return {
+          success: false,
+          output: '',
+          error: error instanceof Error ? error.message : String(error),
+        };
+      }
+    },
+  };
+
+  const cronDelete: Tool = {
+    name: 'cron.delete',
+    description:
+      'Delete a cron job by name. Stops the job immediately. Only works for runtime-created jobs; config-defined jobs will reappear on restart.',
+    inputSchema: {
+      type: 'object',
+      properties: {
+        name: {
+          type: 'string',
+          description: 'Name of the cron job to delete',
+        },
+      },
+      required: ['name'],
+    },
+    execute: async (rawArgs: unknown): Promise<ToolResult> => {
+      const args = rawArgs as { name: string };
+
+      try {
+        const removed = scheduler.removeJob(args.name);
+        if (!removed) {
+          return {
+            success: false,
+            output: '',
+            error: `Cron job "${args.name}" not found. Available: ${scheduler.getJobNames().join(', ')}`,
+          };
+        }
+
+        return {
+          success: true,
+          output: `Cron job "${args.name}" deleted and stopped.`,
+        };
+      } catch (error) {
+        return {
+          success: false,
+          output: '',
+          error: error instanceof Error ? error.message : String(error),
+        };
+      }
+    },
+  };
+
+  return [cronList, cronTrigger, cronCreate, cronDelete];
 }

src/tools/policy.ts

@@ -34,6 +34,10 @@ const PROFILE_TOOLS: Record<ToolProfile, Set<string>> = {
     'drive.read',
     'tasks.lists',
     'tasks.list',
+    'cron.list',
+    'cron.trigger',
+    'cron.create',
+    'cron.delete',
   ]),
   coding: new Set([
     'file.read',
@@ -58,6 +62,10 @@ const PROFILE_TOOLS: Record<ToolProfile, Set<string>> = {
     'drive.read',
     'tasks.lists',
     'tasks.list',
+    'cron.list',
+    'cron.trigger',
+    'cron.create',
+    'cron.delete',
     'file.write',
     'file.edit',
     'file.patch',
@@ -90,6 +98,7 @@ export const TOOL_GROUPS: Record<string, string[]> = {
   'group:gdocs': ['docs.list', 'docs.search', 'docs.read'],
   'group:gdrive': ['drive.list', 'drive.search', 'drive.read'],
   'group:gtasks': ['tasks.lists', 'tasks.list'],
+  'group:cron': ['cron.list', 'cron.trigger', 'cron.create', 'cron.delete'],
 };
 
 /** Expand group references in a list of tool names/patterns. */