fix: prevent agents from duplicating delegated subagent work

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>

src/agents/sisyphus/default.ts

@@ -21,6 +21,7 @@ import {
   buildAntiPatternsSection,
   buildParallelDelegationSection,
   buildNonClaudePlannerSection,
+  buildAntiDuplicationSection,
   categorizeTools,
 } from "../dynamic-agent-prompt-builder";
 
@@ -319,7 +320,7 @@ task(subagent_type="explore", run_in_background=true, load_skills=[], descriptio
 // Reference Grep (external)
 task(subagent_type="librarian", run_in_background=true, load_skills=[], description="Find JWT security docs", prompt="I'm implementing JWT auth and need current security best practices to choose token storage (httpOnly cookies vs localStorage) and set expiration policy. Find: OWASP auth guidelines, recommended token lifetimes, refresh token rotation strategies, common JWT vulnerabilities. Skip 'what is JWT' tutorials — production security guidance only.")
 task(subagent_type="librarian", run_in_background=true, load_skills=[], description="Find Express auth patterns", prompt="I'm building Express auth middleware and need production-quality patterns to structure my middleware chain. Find how established Express apps (1000+ stars) handle: middleware ordering, token refresh, role-based access control, auth error propagation. Skip basic tutorials — I need battle-tested patterns with proper error handling.")
-// Continue working immediately. System notifies on completion — collect with background_output then.
+// Continue only with non-overlapping work. If none exists, end your response and wait for completion.
 
 // WRONG: Sequential or blocking
 result = task(..., run_in_background=false)  // Never wait synchronously for explore/librarian
@@ -327,11 +328,14 @@ result = task(..., run_in_background=false)  // Never wait synchronously for exp
 
 ### Background Result Collection:
 1. Launch parallel agents → receive task_ids
-2. If you have DIFFERENT independent work → do it now
-3. Otherwise → **END YOUR RESPONSE.**
-4. System sends \`<system-reminder>\` on completion → triggers your next turn
-5. Collect via \`background_output(task_id="...")\`
-6. Cleanup: Cancel disposable tasks individually via \`background_cancel(taskId="...")\`
+2. Continue only with non-overlapping work
+   - If you have DIFFERENT independent work → do it now
+   - Otherwise → **END YOUR RESPONSE.**
+3. System sends \`<system-reminder>\` on completion → triggers your next turn
+4. Collect via \`background_output(task_id="...")\`
+5. Cleanup: Cancel disposable tasks individually via \`background_cancel(taskId="...")\`
+
+${buildAntiDuplicationSection()}
 
 ### Search Stop Conditions
 

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

@@ -37,6 +37,7 @@ import {
   buildOracleSection,
   buildHardBlocksSection,
   buildAntiPatternsSection,
+  buildAntiDuplicationSection,
   buildNonClaudePlannerSection,
   categorizeTools,
 } from "../dynamic-agent-prompt-builder";
@@ -233,7 +234,7 @@ ${librarianSection}
 <tool_method>
 - Fire 2-5 explore/librarian agents in parallel for any non-trivial codebase question.
 - Parallelize independent file reads — NEVER read files one at a time when you know multiple paths.
-- When delegating AND doing direct work: do both simultaneously.
+- When delegating AND doing direct work: do only non-overlapping work simultaneously.
 </tool_method>
 
 Explore and Librarian agents are background grep — always \`run_in_background=true\`, always parallel.
@@ -246,11 +247,14 @@ Each agent prompt should include:
 
 Background result collection:
 1. Launch parallel agents → receive task_ids
-2. If you have DIFFERENT independent work → do it now
-3. Otherwise → **END YOUR RESPONSE.**
-4. System sends \`<system-reminder>\` on completion → triggers your next turn
-5. Collect via \`background_output(task_id="...")\`
-6. Cancel disposable tasks individually via \`background_cancel(taskId="...")\`
+2. Continue only with non-overlapping work
+   - If you have DIFFERENT independent work → do it now
+   - Otherwise → **END YOUR RESPONSE.**
+3. System sends \`<system-reminder>\` on completion → triggers your next turn
+4. Collect via \`background_output(task_id="...")\`
+5. Cancel disposable tasks individually via \`background_cancel(taskId="...")\`
+
+${buildAntiDuplicationSection()}
 
 Stop searching when: you have enough context, same info repeating, 2 iterations with no new data, or direct answer found.
 </explore>`;