refactor(task): align continuation ids with task_id

src/agents/atlas/default-prompt-sections.ts

@@ -150,16 +150,16 @@ task(
 
 ### 3.5 Handle Failures (USE RESUME)
 
-**CRITICAL: When re-delegating, ALWAYS use \`session_id\` parameter.**
+**CRITICAL: When re-delegating, ALWAYS use \`task_id\` parameter.**
 
-Every \`task()\` output includes a session_id. STORE IT.
+Every \`task()\` output includes a task_id. STORE IT.
 
 If task fails:
 1. Identify what went wrong
 2. **Resume the SAME session** - subagent has full context already:
     \`\`\`typescript
     task(
-      session_id="ses_xyz789",  // Session from failed task
+      task_id="ses_xyz789",  // Task ID from failed task
       load_skills=[...],
       prompt="FAILED: {error}. Fix by: {specific instruction}"
     )
@@ -167,7 +167,7 @@ If task fails:
 3. Maximum 3 retry attempts with the SAME session
 4. If blocked after 3 attempts: Document and continue to independent tasks
 
-**Why session_id is MANDATORY for failures:**
+**Why task_id is MANDATORY for failures:**
 - Subagent already read all files, knows the context
 - No repeated exploration = 70%+ token savings
 - Subagent knows what approaches already failed
@@ -292,6 +292,6 @@ export const DEFAULT_ATLAS_CRITICAL_RULES = `<critical_overrides>
 - Pass inherited wisdom to every subagent
 - Parallelize independent tasks
 - Verify with your own tools
-- **Store session_id from every delegation output**
-- **Use \`session_id="{session_id}"\` for retries, fixes, and follow-ups**
+- **Store task_id from every delegation output**
+- **Use \`task_id="{task_id}"\` for retries, fixes, and follow-ups**
 </critical_overrides>`

src/agents/atlas/gemini-prompt-sections.ts

@@ -164,10 +164,10 @@ Count remaining **top-level task** checkboxes. Ignore nested verification/eviden
 
 ### 3.5 Handle Failures
 
-**CRITICAL: Use \`session_id\` for retries.**
+**CRITICAL: Use \`task_id\` for retries.**
 
 \`\`\`typescript
-task(session_id="ses_xyz789", load_skills=[...], prompt="FAILED: {error}. Fix by: {instruction}")
+task(task_id="ses_xyz789", load_skills=[...], prompt="FAILED: {error}. Fix by: {instruction}")
 \`\`\`
 
 - Maximum 3 retries per task

src/agents/atlas/gpt-prompt-sections.ts

@@ -169,10 +169,10 @@ Count remaining **top-level task** checkboxes. Ignore nested verification/eviden
 
 ### 3.5 Handle Failures
 
-**CRITICAL: Use \`session_id\` for retries.**
+**CRITICAL: Use \`task_id\` for retries.**
 
 \`\`\`typescript
-task(session_id="ses_xyz789", load_skills=[...], prompt="FAILED: {error}. Fix by: {instruction}")
+task(task_id="ses_xyz789", load_skills=[...], prompt="FAILED: {error}. Fix by: {instruction}")
 \`\`\`
 
 - Maximum 3 retries per task

src/agents/dynamic-agent-core-sections.ts

@@ -182,7 +182,7 @@ Multi-step task? **ALWAYS consult Plan Agent first.** Do NOT start implementatio
 
 - Single-file fix or trivial change → proceed directly
 - Anything else (2+ steps, unclear scope, architecture) → \`task(subagent_type="plan", ...)\` FIRST
-- Use \`session_id\` to resume the same Plan Agent - ask follow-up questions aggressively
+- Use \`task_id\` to resume the same Plan Agent - ask follow-up questions aggressively
 - If ANY part of the task is ambiguous, ask Plan Agent before guessing
 
 Plan Agent returns a structured work breakdown with parallel execution opportunities. Follow it.`

src/agents/hephaestus/gpt-5-3-codex.ts

@@ -409,9 +409,9 @@ After delegation, ALWAYS verify: works as expected? follows codebase pattern? MU
 
 Every \`task()\` output includes a session_id. **USE IT for follow-ups.**
 
-- **Task failed/incomplete** - \`session_id="{id}", prompt="Fix: {error}"\`
-- **Follow-up on result** - \`session_id="{id}", prompt="Also: {question}"\`
-- **Verification failed** - \`session_id="{id}", prompt="Failed: {error}. Fix."\`
+- **Task failed/incomplete** - \`task_id="{id}", prompt="Fix: {error}"\`
+- **Follow-up on result** - \`task_id="{id}", prompt="Also: {question}"\`
+- **Verification failed** - \`task_id="{id}", prompt="Failed: {error}. Fix."\`
 
 ${
   oracleSection

src/agents/hephaestus/gpt-5-4.ts

@@ -312,10 +312,10 @@ Every delegation prompt needs these 6 sections:
 After delegation, verify by reading every file the subagent touched. Check: works as expected? follows codebase pattern? Do not trust self-reports.
 
 <session_continuity>
-Every \`task()\` returns a session_id. Use it for all follow-ups:
-- Task failed/incomplete: \`session_id="{id}", prompt="Fix: {error}"\`
-- Follow-up on result: \`session_id="{id}", prompt="Also: {question}"\`
-- Verification failed: \`session_id="{id}", prompt="Failed: {error}. Fix."\`
+Every \`task()\` returns a task_id. Use it for all follow-ups:
+- Task failed/incomplete: \`task_id="{id}", prompt="Fix: {error}"\`
+- Follow-up on result: \`task_id="{id}", prompt="Also: {question}"\`
+- Verification failed: \`task_id="{id}", prompt="Failed: {error}. Fix."\`
 
 This preserves full context, avoids repeated exploration, saves 70%+ tokens.
 </session_continuity>

src/agents/hephaestus/gpt.ts

@@ -277,11 +277,11 @@ After delegation, ALWAYS verify: works as expected? follows codebase pattern? MU
 
 ### Session Continuity
 
-Every \`task()\` output includes a session_id. **USE IT for follow-ups.**
+Every \`task()\` output includes a task_id. **USE IT for follow-ups.**
 
-- **Task failed/incomplete** - \`session_id="{id}", prompt="Fix: {error}"\`
-- **Follow-up on result** - \`session_id="{id}", prompt="Also: {question}"\`
-- **Verification failed** - \`session_id="{id}", prompt="Failed: {error}. Fix."\`
+- **Task failed/incomplete** - \`task_id="{id}", prompt="Fix: {error}"\`
+- **Follow-up on result** - \`task_id="{id}", prompt="Also: {question}"\`
+- **Verification failed** - \`task_id="{id}", prompt="Failed: {error}. Fix."\`
 
 ${
   oracleSection

src/agents/sisyphus.ts

@@ -317,15 +317,15 @@ AFTER THE WORK YOU DELEGATED SEEMS DONE, ALWAYS VERIFY THE RESULTS AS FOLLOWING:
 
 ### Session Continuity (MANDATORY)
 
-Every \`task()\` output includes a session_id. **USE IT.**
+Every \`task()\` output includes a task_id. **USE IT.**
 
 **ALWAYS continue when:**
-- Task failed/incomplete → \`session_id=\"{session_id}\", prompt=\"Fix: {specific error}\"\`
-- Follow-up question on result → \`session_id=\"{session_id}\", prompt=\"Also: {question}\"\`
-- Multi-turn with same agent → \`session_id=\"{session_id}\"\` - NEVER start fresh
-- Verification failed → \`session_id=\"{session_id}\", prompt=\"Failed verification: {error}. Fix.\"\`
+- Task failed/incomplete → \`task_id=\"{task_id}\", prompt=\"Fix: {specific error}\"\`
+- Follow-up question on result → \`task_id=\"{task_id}\", prompt=\"Also: {question}\"\`
+- Multi-turn with same agent → \`task_id=\"{task_id}\"\` - NEVER start fresh
+- Verification failed → \`task_id=\"{task_id}\", prompt=\"Failed verification: {error}. Fix.\"\`
 
-**Why session_id is CRITICAL:**
+**Why task_id is CRITICAL:**
 - Subagent has FULL conversation context preserved
 - No repeated file reads, exploration, or setup
 - Saves 70%+ tokens on follow-ups
@@ -336,10 +336,10 @@ Every \`task()\` output includes a session_id. **USE IT.**
 task(category="quick", load_skills=[], run_in_background=false, description="Fix type error", prompt="Fix the type error in auth.ts...")
 
 // CORRECT: Resume preserves everything
-task(session_id="ses_abc123", load_skills=[], run_in_background=false, description="Fix type error", prompt="Fix: Type error on line 42")
+task(task_id="ses_abc123", load_skills=[], run_in_background=false, description="Fix type error", prompt="Fix: Type error on line 42")
 \`\`\`
 
-**After EVERY delegation, STORE the session_id for potential continuation.**
+**After EVERY delegation, STORE the task_id for potential continuation.**
 
 ### Code Changes:
 - Match existing patterns (if codebase is disciplined)

src/agents/sisyphus/default.ts

@@ -389,15 +389,15 @@ AFTER THE WORK YOU DELEGATED SEEMS DONE, ALWAYS VERIFY THE RESULTS AS FOLLOWING:
 
 ### Session Continuity (MANDATORY)
 
-Every \`task()\` output includes a session_id. **USE IT.**
+Every \`task()\` output includes a task_id. **USE IT.**
 
 **ALWAYS continue when:**
-- Task failed/incomplete → \`session_id="{session_id}", prompt="Fix: {specific error}"\`
-- Follow-up question on result → \`session_id="{session_id}", prompt="Also: {question}"\`
-- Multi-turn with same agent → \`session_id="{session_id}"\` - NEVER start fresh
-- Verification failed → \`session_id="{session_id}", prompt="Failed verification: {error}. Fix."\`
+- Task failed/incomplete → \`task_id="{task_id}", prompt="Fix: {specific error}"\`
+- Follow-up question on result → \`task_id="{task_id}", prompt="Also: {question}"\`
+- Multi-turn with same agent → \`task_id="{task_id}"\` - NEVER start fresh
+- Verification failed → \`task_id="{task_id}", prompt="Failed verification: {error}. Fix."\`
 
-**Why session_id is CRITICAL:**
+**Why task_id is CRITICAL:**
 - Subagent has FULL conversation context preserved
 - No repeated file reads, exploration, or setup
 - Saves 70%+ tokens on follow-ups
@@ -408,10 +408,10 @@ Every \`task()\` output includes a session_id. **USE IT.**
 task(category="quick", load_skills=[], run_in_background=false, description="Fix type error", prompt="Fix the type error in auth.ts...")
 
 // CORRECT: Resume preserves everything
-task(session_id="ses_abc123", load_skills=[], run_in_background=false, description="Fix type error", prompt="Fix: Type error on line 42")
+task(task_id="ses_abc123", load_skills=[], run_in_background=false, description="Fix type error", prompt="Fix: Type error on line 42")
 \`\`\`
 
-**After EVERY delegation, STORE the session_id for potential continuation.**
+**After EVERY delegation, STORE the task_id for potential continuation.**
 
 ### Code Changes:
 - Match existing patterns (if codebase is disciplined)

src/agents/sisyphus/gpt-5-4.ts

@@ -387,10 +387,10 @@ Post-delegation: delegation never substitutes for verification. Always run \`<ve
 
 ### Session continuity
 
-Every \`task()\` returns a session_id. Use it for all follow-ups:
-- Failed/incomplete → \`session_id="{id}", prompt="Fix: {specific error}"\`
-- Follow-up → \`session_id="{id}", prompt="Also: {question}"\`
-- Multi-turn → always \`session_id\`, never start fresh
+Every \`task()\` returns a task_id. Use it for all follow-ups:
+- Failed/incomplete → \`task_id="{id}", prompt="Fix: {specific error}"\`
+- Follow-up → \`task_id="{id}", prompt="Also: {question}"\`
+- Multi-turn → always \`task_id\`, never start fresh
 
 This preserves full context, avoids repeated exploration, saves 70%+ tokens.
 

src/features/tool-metadata-store/task-metadata-contract.test.ts

@@ -18,7 +18,7 @@ describe("buildTaskMetadataBlock", () => {
     // given
     const link = {
       sessionId: "ses_bg_123",
-      taskId: "bg_123",
+      taskId: "ses_bg_123",
       backgroundTaskId: "bg_123",
       agent: "explore",
       category: "quick",
@@ -29,7 +29,7 @@ describe("buildTaskMetadataBlock", () => {
 
     // then
     expect(block).toBe(
-      "<task_metadata>\nsession_id: ses_bg_123\ntask_id: bg_123\nbackground_task_id: bg_123\nsubagent: explore\ncategory: quick\n</task_metadata>"
+      "<task_metadata>\nsession_id: ses_bg_123\ntask_id: ses_bg_123\nbackground_task_id: bg_123\nsubagent: explore\ncategory: quick\n</task_metadata>"
     )
   })
 })

src/hooks/atlas/verification-reminders.ts

@@ -29,7 +29,7 @@ Your completion will NOT be recorded until you complete ALL of the following:
 
 If anything fails while closing this out, resume the same session immediately:
 \`\`\`typescript
-task(session_id="${sessionId}", load_skills=[], prompt="fix: checkbox not recorded correctly")
+task(task_id="${sessionId}", load_skills=[], prompt="fix: checkbox not recorded correctly")
 \`\`\`
 
 **Your completion is NOT tracked until the checkbox is marked in the plan file.**
@@ -47,7 +47,7 @@ ${VERIFICATION_REMINDER}
 
 **If ANY verification fails, use this immediately:**
 \`\`\`
-task(session_id="${sessionId}", load_skills=[], prompt="fix: [describe the specific failure]")
+task(task_id="${sessionId}", load_skills=[], prompt="fix: [describe the specific failure]")
 \`\`\`
 
 ${buildReuseHint(sessionId)}`

src/hooks/keyword-detector/ultrawork/default.ts

@@ -115,15 +115,15 @@ task(subagent_type="plan", load_skills=[], run_in_background=false, prompt="<gat
 
 ### SESSION CONTINUITY WITH PLAN AGENT (CRITICAL)
 
-**Plan agent returns a session_id. USE IT for follow-up interactions.**
+**Plan agent returns a task_id. USE IT for follow-up interactions.**
 
 | Scenario | Action |
 |----------|--------|
-| Plan agent asks clarifying questions | \`task(session_id="{returned_session_id}", load_skills=[], run_in_background=false, prompt="<your answer>")\` |
-| Need to refine the plan | \`task(session_id="{returned_session_id}", load_skills=[], run_in_background=false, prompt="Please adjust: <feedback>")\` |
-| Plan needs more detail | \`task(session_id="{returned_session_id}", load_skills=[], run_in_background=false, prompt="Add more detail to Task N")\` |
+| Plan agent asks clarifying questions | \`task(task_id="{returned_task_id}", load_skills=[], run_in_background=false, prompt="<your answer>")\` |
+| Need to refine the plan | \`task(task_id="{returned_task_id}", load_skills=[], run_in_background=false, prompt="Please adjust: <feedback>")\` |
+| Plan needs more detail | \`task(task_id="{returned_task_id}", load_skills=[], run_in_background=false, prompt="Add more detail to Task N")\` |
 
-**WHY SESSION_ID IS CRITICAL:**
+**WHY TASK_ID IS CRITICAL:**
 - Plan agent retains FULL conversation context
 - No repeated exploration or context gathering
 - Saves 70%+ tokens on follow-ups
@@ -134,7 +134,7 @@ task(subagent_type="plan", load_skills=[], run_in_background=false, prompt="<gat
 task(subagent_type="plan", load_skills=[], run_in_background=false, prompt="Here's more info...")
 
 // CORRECT: Resume preserves everything
-task(session_id="ses_abc123", load_skills=[], run_in_background=false, prompt="Here's my answer to your question: ...")
+task(task_id="ses_abc123", load_skills=[], run_in_background=false, prompt="Here's my answer to your question: ...")
 \`\`\`
 
 **FAILURE TO CALL PLAN AGENT = INCOMPLETE WORK.**

src/hooks/keyword-detector/ultrawork/gemini.ts

@@ -161,13 +161,13 @@ task(subagent_type="plan", load_skills=[], run_in_background=false, prompt="<gat
 
 ### SESSION CONTINUITY WITH PLAN AGENT (CRITICAL)
 
-**Plan agent returns a session_id. USE IT for follow-up interactions.**
+**Plan agent returns a task_id. USE IT for follow-up interactions.**
 
 | Scenario | Action |
 |----------|--------|
-| Plan agent asks clarifying questions | \`task(session_id="{returned_session_id}", load_skills=[], run_in_background=false, prompt="<your answer>")\` |
-| Need to refine the plan | \`task(session_id="{returned_session_id}", load_skills=[], run_in_background=false, prompt="Please adjust: <feedback>")\` |
-| Plan needs more detail | \`task(session_id="{returned_session_id}", load_skills=[], run_in_background=false, prompt="Add more detail to Task N")\` |
+| Plan agent asks clarifying questions | \`task(task_id="{returned_task_id}", load_skills=[], run_in_background=false, prompt="<your answer>")\` |
+| Need to refine the plan | \`task(task_id="{returned_task_id}", load_skills=[], run_in_background=false, prompt="Please adjust: <feedback>")\` |
+| Plan needs more detail | \`task(task_id="{returned_task_id}", load_skills=[], run_in_background=false, prompt="Add more detail to Task N")\` |
 
 **FAILURE TO CALL PLAN AGENT = INCOMPLETE WORK.**
 

src/hooks/task-resume-info/hook.ts

@@ -12,12 +12,13 @@ export function createTaskResumeInfoHook() {
     if (outputText.startsWith("Error:") || outputText.startsWith("Failed")) return
     if (outputText.includes("\nto continue:")) return
 
-    const sessionId = extractTaskLink(output.metadata, outputText).sessionId
-    if (!sessionId) return
+    const link = extractTaskLink(output.metadata, outputText)
+    const taskId = link.taskId ?? link.sessionId
+    if (!taskId) return
 
     output.output =
       outputText.trimEnd() +
-      `\n\nto continue: task(session_id="${sessionId}", load_skills=[], run_in_background=false, prompt="...")`
+      `\n\nto continue: task(task_id="${taskId}", load_skills=[], run_in_background=false, prompt="...")`
   }
 
   return {

src/hooks/task-resume-info/index.test.ts

@@ -60,6 +60,7 @@ describe("createTaskResumeInfoHook", () => {
         await afterHook(input, output)
 
         expect(output.output).toContain("to continue:")
+        expect(output.output).toContain('task(task_id="ses_abc123"')
         expect(output.output).toContain("ses_abc123")
       })
 
@@ -74,6 +75,7 @@ describe("createTaskResumeInfoHook", () => {
         await afterHook(input, output)
 
         expect(output.output).toContain("run_in_background=false")
+        expect(output.output).toContain('task_id="ses_abc123"')
       })
     })
   })
@@ -120,7 +122,7 @@ describe("createTaskResumeInfoHook", () => {
         const output = {
           title: "task",
           output:
-            'Done.\nSession ID: ses_abc123\nto continue: task(session_id="ses_abc123", load_skills=[], prompt="...")',
+            'Done.\nSession ID: ses_abc123\nto continue: task(task_id="ses_abc123", load_skills=[], run_in_background=false, prompt="...")',
           metadata: {},
         }
 

src/plugin/tool-execute-before.test.ts

@@ -161,6 +161,23 @@ describe("createToolExecuteBeforeHandler", () => {
       expect(output.args.subagent_type).toBe("explore")
     })
 
+    test("normalizes task_id into the canonical resume argument", async () => {
+      //#given
+      const ctx = createCtxWithSessionMessages([
+        { info: { role: "assistant", agent: "oracle" } },
+      ])
+      const handler = createToolExecuteBeforeHandler({ ctx, hooks: emptyHooks })
+      const input = { tool: "task", sessionID: "ses_123", callID: "call_1" }
+      const output = { args: { task_id: "ses_resume_123", description: "Continue task", prompt: "fix it" } as Record<string, unknown> }
+
+      //#when
+      await handler(input, output)
+
+      //#then
+      expect(output.args.task_id).toBe("ses_resume_123")
+      expect(output.args.subagent_type).toBe("oracle")
+    })
+
     test("falls back to 'continue' when session has no agent info", async () => {
       //#given
       const ctx = createCtxWithSessionMessages([

src/plugin/tool-execute-before.ts

@@ -100,12 +100,21 @@ export function createToolExecuteBeforeHandler(args: {
       const argsObject = output.args
       const category = typeof argsObject.category === "string" ? argsObject.category : undefined
       const subagentType = typeof argsObject.subagent_type === "string" ? argsObject.subagent_type : undefined
-      const sessionId = typeof argsObject.session_id === "string" ? argsObject.session_id : undefined
+      const taskId =
+        typeof argsObject.task_id === "string"
+          ? argsObject.task_id
+          : typeof argsObject.session_id === "string"
+            ? argsObject.session_id
+            : undefined
+
+      if (taskId && typeof argsObject.task_id !== "string") {
+        argsObject.task_id = taskId
+      }
 
       if (category) {
         argsObject.subagent_type = "sisyphus-junior"
-      } else if (!subagentType && sessionId) {
-        const resolvedAgent = await resolveSessionAgent(ctx.client, sessionId)
+      } else if (!subagentType && taskId) {
+        const resolvedAgent = await resolveSessionAgent(ctx.client, taskId)
         argsObject.subagent_type = resolvedAgent ?? "continue"
       }
 

src/tools/background-task/create-background-cancel.ts

@@ -59,7 +59,7 @@ export function createBackgroundCancel(manager: BackgroundManager, _client: Back
 
 To continue a cancelled task, use:
 \`\`\`
-task(session_id="<session_id>", prompt="Continue: <your follow-up>")
+task(task_id="<task_id>", prompt="Continue: <your follow-up>")
 \`\`\`
 
 Continuable sessions:

src/tools/delegate-task/background-continuation.ts

@@ -3,6 +3,8 @@ import type { ExecutorContext, ParentContext } from "./executor-types"
 import { publishToolMetadata } from "../../features/tool-metadata-store"
 import { formatDetailedError } from "./error-formatting"
 import { getSessionTools } from "../../shared/session-tools-store"
+import { buildTaskMetadataBlock } from "../../features/tool-metadata-store/task-metadata-contract"
+import { getTaskID } from "./task-id"
 
 export async function executeBackgroundContinuation(
   args: DelegateTaskArgs,
@@ -13,8 +15,13 @@ export async function executeBackgroundContinuation(
   const { manager } = executorCtx
 
   try {
+    const taskID = getTaskID(args)
+    if (!taskID) {
+      throw new Error("task_id is required to continue a background task")
+    }
+
     const task = await manager.resume({
-      sessionId: args.session_id!,
+      sessionId: taskID,
       prompt: args.prompt,
       parentSessionID: parentContext.sessionID,
       parentMessageID: parentContext.messageID,
@@ -31,6 +38,8 @@ export async function executeBackgroundContinuation(
         load_skills: args.load_skills,
         description: args.description,
         run_in_background: args.run_in_background,
+        taskId: task.sessionID,
+        backgroundTaskId: task.id,
         sessionId: task.sessionID,
         command: args.command,
         model: task.model ? { providerID: task.model.providerID, modelID: task.model.modelID } : undefined,
@@ -50,14 +59,17 @@ System notifies on completion. Use \`background_output\` with task_id="${task.id
 
 Do NOT call background_output now. Wait for <system-reminder> notification first.
 
-<task_metadata>
-session_id: ${task.sessionID}
-${task.agent ? `subagent: ${task.agent}\n` : ""}</task_metadata>`
+${buildTaskMetadataBlock({
+      sessionId: task.sessionID,
+      taskId: task.sessionID,
+      backgroundTaskId: task.id,
+      agent: task.agent,
+    })}`
   } catch (error) {
     return formatDetailedError(error, {
       operation: "Continue background task",
       args,
-      sessionID: args.session_id,
+      sessionID: getTaskID(args),
     })
   }
 }

src/tools/delegate-task/background-task.test.ts

@@ -104,11 +104,14 @@ describeFn("executeBackgroundTask output/session metadata compatibility", () =>
     //#then - output and metadata should include canonical session linkage
     expectFn(result).toContain("<task_metadata>")
     expectFn(result).toContain("session_id: ses_sub_123")
-    expectFn(result).toContain("task_id: bg_resolved")
+    expectFn(result).toContain("task_id: ses_sub_123")
     expectFn(result).toContain("background_task_id: bg_resolved")
+    expectFn(result).toContain("subagent: explore")
     expectFn(result).toContain("Background Task ID: bg_resolved")
     expectFn(metadataCalls).toHaveLength(1)
     expectFn(metadataCalls[0].metadata.sessionId).toBe("ses_sub_123")
+    expectFn(metadataCalls[0].metadata.taskId).toBe("ses_sub_123")
+    expectFn(metadataCalls[0].metadata.backgroundTaskId).toBe("bg_resolved")
   })
 
   testFn("captures late-resolved session id and emits synced metadata", async () => {
@@ -152,10 +155,12 @@ describeFn("executeBackgroundTask output/session metadata compatibility", () =>
 
     //#then - late session id still propagates to task metadata contract
     expectFn(result).toContain("session_id: ses_late_123")
-    expectFn(result).toContain("task_id: bg_late")
+    expectFn(result).toContain("task_id: ses_late_123")
     expectFn(result).toContain("background_task_id: bg_late")
     expectFn(metadataCalls).toHaveLength(1)
     expectFn(metadataCalls[0].metadata.sessionId).toBe("ses_late_123")
+    expectFn(metadataCalls[0].metadata.taskId).toBe("ses_late_123")
+    expectFn(metadataCalls[0].metadata.backgroundTaskId).toBe("bg_late")
   })
 
   testFn("passes question-deny session permission when launching delegate task", async () => {

src/tools/delegate-task/background-task.ts

@@ -10,6 +10,7 @@ import { SessionCategoryRegistry } from "../../shared/session-category-registry"
 import { QUESTION_DENIED_SESSION_PERMISSION } from "../../shared/question-denied-session-permission"
 import { setSessionFallbackChain } from "../../hooks/model-fallback/hook"
 import { stripAgentListSortPrefix } from "../../shared/agent-display-names"
+import { buildTaskMetadataBlock } from "../../features/tool-metadata-store/task-metadata-contract"
 
 function continueSessionSetup(args: {
   taskID: string
@@ -125,6 +126,8 @@ export async function executeBackgroundTask(
       description: args.description,
       run_in_background: args.run_in_background,
       command: args.command,
+      ...(sessionId ? { taskId: sessionId } : {}),
+      backgroundTaskId: task.id,
       ...(sessionId ? { sessionId } : {}),
       ...(categoryModel ? { model: { providerID: categoryModel.providerID, modelID: categoryModel.modelID } } : {}),
     }
@@ -136,7 +139,13 @@ export async function executeBackgroundTask(
     await publishToolMetadata(ctx, unstableMeta)
 
     const taskMetadataBlock = sessionId
-      ? `\n\n<task_metadata>\nsession_id: ${sessionId}\ntask_id: ${task.id}\nbackground_task_id: ${task.id}\n</task_metadata>`
+      ? `\n\n${buildTaskMetadataBlock({
+        sessionId,
+        taskId: sessionId,
+        backgroundTaskId: task.id,
+        agent: task.agent,
+        category: args.category,
+      })}`
       : ""
 
     return `Background task launched.

src/tools/delegate-task/error-formatting.ts

@@ -1,4 +1,5 @@
 import type { DelegateTaskArgs } from "./types"
+import { getTaskID } from "./task-id"
 
 /**
  * Context for error formatting.
@@ -35,8 +36,9 @@ export function formatDetailedError(error: unknown, ctx: ErrorContext): string {
     lines.push(`- subagent_type: ${ctx.args.subagent_type ?? "(none)"}`)
     lines.push(`- run_in_background: ${ctx.args.run_in_background}`)
     lines.push(`- load_skills: [${ctx.args.load_skills?.join(", ") ?? ""}]`)
-    if (ctx.args.session_id) {
-      lines.push(`- session_id: ${ctx.args.session_id}`)
+    const taskID = getTaskID(ctx.args)
+    if (taskID) {
+      lines.push(`- task_id: ${taskID}`)
     }
   }
 

src/tools/delegate-task/sync-continuation.ts

@@ -12,6 +12,8 @@ import { syncContinuationDeps, type SyncContinuationDeps } from "./sync-continua
 import { setSessionTools } from "../../shared/session-tools-store"
 import { normalizeSDKResponse } from "../../shared"
 import { buildTaskPrompt } from "./prompt-builder"
+import { buildTaskMetadataBlock } from "../../features/tool-metadata-store/task-metadata-contract"
+import { getTaskID } from "./task-id"
 
 export async function executeSyncContinuation(
   args: DelegateTaskArgs,
@@ -21,7 +23,11 @@ export async function executeSyncContinuation(
 ): Promise<string> {
   const { client, syncPollTimeoutMs, sisyphusAgentConfig } = executorCtx
   const toastManager = getTaskToastManager()
-  const taskId = `resume_sync_${args.session_id!.slice(0, 8)}`
+  const continuationID = getTaskID(args)
+  if (!continuationID) {
+    throw new Error("task_id is required to continue a sync task")
+  }
+  const taskId = `resume_sync_${continuationID.slice(0, 8)}`
   const startTime = new Date()
 
   if (toastManager) {
@@ -42,7 +48,7 @@ export async function executeSyncContinuation(
 
   try {
     try {
-      const messagesResp = await client.session.messages({ path: { id: args.session_id! } })
+      const messagesResp = await client.session.messages({ path: { id: continuationID } })
       const messages = normalizeSDKResponse(messagesResp, [] as SessionMessage[])
       anchorMessageCount = messages.length
       for (let i = messages.length - 1; i >= 0; i--) {
@@ -55,7 +61,7 @@ export async function executeSyncContinuation(
         }
       }
     } catch {
-      const resumeMessageDir = getMessageDir(args.session_id!)
+      const resumeMessageDir = getMessageDir(continuationID)
       const resumeMessage = resumeMessageDir ? findNearestMessageWithFields(resumeMessageDir) : null
       resumeAgent = resumeMessage?.agent
       resumeModel = resumeMessage?.model?.providerID && resumeMessage?.model?.modelID
@@ -71,7 +77,8 @@ export async function executeSyncContinuation(
         load_skills: args.load_skills,
         description: args.description,
         run_in_background: args.run_in_background,
-        sessionId: args.session_id,
+        taskId: continuationID,
+        sessionId: continuationID,
         sync: true,
         command: args.command,
         model: resumeModel,
@@ -88,10 +95,10 @@ export async function executeSyncContinuation(
       question: false,
       ...(resumeAgent ? getAgentToolRestrictions(resumeAgent) : {}),
     }
-    setSessionTools(args.session_id!, tools)
+    setSessionTools(continuationID, tools)
 
     await promptWithModelSuggestionRetry(client, {
-      path: { id: args.session_id! },
+      path: { id: continuationID },
       body: {
         ...(resumeAgent !== undefined ? { agent: resumeAgent } : {}),
         ...(resumeModel !== undefined ? { model: resumeModel } : {}),
@@ -105,12 +112,12 @@ export async function executeSyncContinuation(
        toastManager.removeTask(taskId)
      }
      const errorMessage = promptError instanceof Error ? promptError.message : String(promptError)
-     return `Failed to send continuation prompt: ${errorMessage}\n\nSession ID: ${args.session_id}`
+     return `Failed to send continuation prompt: ${errorMessage}\n\nTask ID: ${continuationID}`
    }
 
     try {
       const pollError = await deps.pollSyncSession(ctx, client, {
-        sessionID: args.session_id!,
+        sessionID: continuationID,
         agentToUse: resumeAgent ?? "continue",
         toastManager,
         taskId,
@@ -120,7 +127,7 @@ export async function executeSyncContinuation(
         return pollError
       }
 
-      const result = await deps.fetchSyncResult(client, args.session_id!, anchorMessageCount)
+      const result = await deps.fetchSyncResult(client, continuationID, anchorMessageCount)
       if (!result.ok) {
         return result.error
       }
@@ -133,9 +140,11 @@ export async function executeSyncContinuation(
 
 ${result.textContent || "(No text output)"}
 
-<task_metadata>
-session_id: ${args.session_id}
-${resumeAgent ? `subagent: ${resumeAgent}\n` : ""}</task_metadata>`
+${buildTaskMetadataBlock({
+        sessionId: continuationID,
+        taskId: continuationID,
+        agent: resumeAgent,
+      })}`
    } finally {
      if (toastManager) {
        toastManager.removeTask(taskId)

src/tools/delegate-task/sync-task.ts

@@ -11,6 +11,7 @@ import { formatDetailedError } from "./error-formatting"
 import { syncTaskDeps, type SyncTaskDeps } from "./sync-task-deps"
 import { setSessionFallbackChain, clearSessionFallbackChain } from "../../hooks/model-fallback/hook"
 import { retrySyncPromptWithFallbacks } from "./sync-task-fallback"
+import { buildTaskMetadataBlock } from "../../features/tool-metadata-store/task-metadata-contract"
 
 export async function executeSyncTask(
   args: DelegateTaskArgs,
@@ -122,6 +123,7 @@ export async function executeSyncTask(
         load_skills: args.load_skills,
         description: args.description,
         run_in_background: args.run_in_background,
+        taskId: sessionID,
         sessionId: sessionID,
         sync: true,
         spawnDepth: spawnContext.childDepth,
@@ -210,9 +212,12 @@ Agent: ${agentToUse}${args.category ? ` (category: ${args.category})` : ""}${mod
 
 ${result.textContent || "(No text output)"}
 
-<task_metadata>
-session_id: ${sessionID}
-</task_metadata>`
+${buildTaskMetadataBlock({
+        sessionId: sessionID,
+        taskId: sessionID,
+        agent: agentToUse,
+        category: args.category,
+      })}`
     } finally {
       if (toastManager && taskId !== undefined) {
         toastManager.removeTask(taskId)

src/tools/delegate-task/task-id.ts

@@ -0,0 +1,5 @@
+import type { DelegateTaskArgs } from "./types"
+
+export function getTaskID(args: Pick<DelegateTaskArgs, "task_id" | "session_id">): string | undefined {
+  return args.task_id ?? args.session_id
+}

src/tools/delegate-task/task-schema.test.ts

@@ -41,6 +41,7 @@ function createDelegateTask(...args: Parameters<typeof import("./tools").createD
 
 		//#then
 		expect(description).toContain("subagent_type: Use specific agent directly")
+		expect(description).toContain("task_id: Existing task to continue")
 		expect(description).not.toContain("sisyphus")
 		expect(description).not.toContain("hephaestus")
 		expect(description).not.toContain("prometheus")

src/tools/delegate-task/tools.ts

@@ -84,13 +84,14 @@ export function createDelegateTask(options: DelegateTaskToolOptions): ToolDefini
   ${categoryList}
   - subagent_type: Use specific agent directly (explore, librarian, oracle, metis, momus)
   - run_in_background: REQUIRED. true=async (returns task_id), false=sync (waits). Use background=true ONLY for parallel exploration with 5+ independent queries.
-  - session_id: Existing Task session to continue (from previous task output). Continues agent with FULL CONTEXT PRESERVED - saves tokens, maintains continuity.
+  - task_id: Existing task to continue (from previous task output). Continues the same subagent session with FULL CONTEXT PRESERVED.
+  - session_id: Deprecated alias for task_id. Accepted for backward compatibility.
   - command: The command that triggered this task (optional, for slash command tracking).
   
-  **WHEN TO USE session_id:**
-  - Task failed/incomplete → session_id with "fix: [specific issue]"
-  - Need follow-up on previous result → session_id with additional question
-  - Multi-turn conversation with same agent → always session_id instead of new task
+  **WHEN TO USE task_id:**
+  - Task failed/incomplete → task_id with "fix: [specific issue]"
+  - Need follow-up on previous result → task_id with additional question
+  - Multi-turn conversation with same agent → always task_id instead of new task
   
   Prompts MUST be in English.`
 
@@ -103,11 +104,15 @@ export function createDelegateTask(options: DelegateTaskToolOptions): ToolDefini
       run_in_background: tool.schema.boolean().describe("REQUIRED. true=async (returns task_id), false=sync (waits). Use false for task delegation, true ONLY for parallel exploration."),
       category: tool.schema.string().optional().describe(`REQUIRED if subagent_type not provided. Do NOT provide both category and subagent_type.`),
       subagent_type: tool.schema.string().optional().describe("REQUIRED if category not provided. Do NOT provide both category and subagent_type."),
-      session_id: tool.schema.string().optional().describe("Existing Task session to continue"),
+      task_id: tool.schema.string().optional().describe("Existing task to continue. Canonical resume identifier."),
+      session_id: tool.schema.string().optional().describe("Deprecated alias for task_id. Existing task to continue."),
       command: tool.schema.string().optional().describe("The command that triggered this task"),
     },
     async execute(args: DelegateTaskArgs, toolContext) {
       const ctx = toolContext as ToolContextWithMetadata
+      if (!args.task_id && args.session_id) {
+        args.task_id = args.session_id
+      }
 
       if (args.category) {
         if (args.subagent_type && args.subagent_type !== SISYPHUS_JUNIOR_AGENT) {
@@ -158,7 +163,7 @@ export function createDelegateTask(options: DelegateTaskToolOptions): ToolDefini
 
       const parentContext = await resolveParentContext(ctx, options.client)
 
-      if (args.session_id) {
+      if (args.task_id || args.session_id) {
         if (runInBackground) {
           return executeBackgroundContinuation(args, ctx, options, parentContext)
         }

src/tools/delegate-task/types.ts

@@ -14,6 +14,8 @@ export interface DelegateTaskArgs {
   category?: string
   subagent_type?: string
   run_in_background: boolean
+  task_id?: string
+  /** @deprecated Use task_id instead. */
   session_id?: string
   command?: string
   load_skills: string[]