Merge pull request #3436 from code-yeongyu/fix/bug-batch-1

fix: numeric skill names, ZWSP agent lookups, ultrawork run_in_background

src/features/opencode-skill-loader/async-loader.ts

@@ -87,7 +87,7 @@ export async function loadSkillFromPathAsync(
     const mcpJsonMcp = await loadMcpJsonFromDirAsync(resolvedPath)
     const mcpConfig = mcpJsonMcp || frontmatterMcp
 
-    const baseName = data.name || defaultName
+    const baseName = String(data.name || defaultName)
     const skillName = namePrefix ? `${namePrefix}/${baseName}` : baseName
     const originalDescription = data.description || ""
     const isOpencodeSource = scope === "opencode" || scope === "opencode-project"

src/features/opencode-skill-loader/loaded-skill-from-path.ts

@@ -25,7 +25,7 @@ export async function loadSkillFromPath(options: {
     const mcpJsonMcp = await loadMcpJsonFromDir(options.resolvedPath)
     const mcpConfig = mcpJsonMcp || frontmatterMcp
 
-    const baseName = data.name || options.defaultName
+    const baseName = String(data.name || options.defaultName)
     const skillName = namePrefix ? `${namePrefix}/${baseName}` : baseName
     const originalDescription = data.description || ""
     const isOpencodeSource = options.scope === "opencode" || options.scope === "opencode-project"

src/hooks/compaction-context-injector/recovery-prompt-config.ts

@@ -1,3 +1,4 @@
+import { stripInvisibleAgentCharacters } from "../../shared/agent-display-names"
 import type { CompactionAgentConfigCheckpoint } from "../../shared/compaction-agent-config-checkpoint"
 
 export type RecoveryPromptConfig = CompactionAgentConfigCheckpoint & {
@@ -66,7 +67,7 @@ export function isPromptConfigRecovered(
   const agentMatches =
     typeof actualAgent === "string" &&
     !isCompactionAgent(actualAgent) &&
-    actualAgent.toLowerCase() === expectedPromptConfig.agent.toLowerCase()
+    stripInvisibleAgentCharacters(actualAgent).toLowerCase() === stripInvisibleAgentCharacters(expectedPromptConfig.agent).toLowerCase()
 
   return (
     agentMatches &&

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

@@ -104,7 +104,7 @@ TELL THE USER WHAT AGENTS YOU WILL LEVERAGE NOW TO SATISFY USER'S REQUEST.
 | Architecture decision needed | MUST call plan agent |
 
 \`\`\`
-task(subagent_type="plan", load_skills=[], prompt="<gathered context + user request>")
+task(subagent_type="plan", load_skills=[], run_in_background=false, prompt="<gathered context + user request>")
 \`\`\`
 
 **WHY PLAN AGENT IS MANDATORY:**
@@ -119,9 +119,9 @@ task(subagent_type="plan", load_skills=[], prompt="<gathered context + user requ
 
 | Scenario | Action |
 |----------|--------|
-| Plan agent asks clarifying questions | \`task(session_id="{returned_session_id}", load_skills=[], prompt="<your answer>")\` |
-| Need to refine the plan | \`task(session_id="{returned_session_id}", load_skills=[], prompt="Please adjust: <feedback>")\` |
-| Plan needs more detail | \`task(session_id="{returned_session_id}", load_skills=[], prompt="Add more detail to Task N")\` |
+| 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")\` |
 
 **WHY SESSION_ID IS CRITICAL:**
 - Plan agent retains FULL conversation context
@@ -131,10 +131,10 @@ task(subagent_type="plan", load_skills=[], prompt="<gathered context + user requ
 
 \`\`\`
 // WRONG: Starting fresh loses all context
-task(subagent_type="plan", load_skills=[], prompt="Here's more info...")
+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=[], prompt="Here's my answer to your question: ...")
+task(session_id="ses_abc123", load_skills=[], run_in_background=false, prompt="Here's my answer to your question: ...")
 \`\`\`
 
 **FAILURE TO CALL PLAN AGENT = INCOMPLETE WORK.**
@@ -149,21 +149,21 @@ task(session_id="ses_abc123", load_skills=[], prompt="Here's my answer to your q
 |-----------|--------|-----|
 | Codebase exploration | task(subagent_type="explore", load_skills=[], run_in_background=true) | Parallel, context-efficient |
 | Documentation lookup | task(subagent_type="librarian", load_skills=[], run_in_background=true) | Specialized knowledge |
-| Planning | task(subagent_type="plan", load_skills=[]) | Parallel task graph + structured TODO list |
-| Hard problem (conventional) | task(subagent_type="oracle", load_skills=[]) | Architecture, debugging, complex logic |
-| Hard problem (non-conventional) | task(category="artistry", load_skills=[...]) | Different approach needed |
-| Implementation | task(category="...", load_skills=[...]) | Domain-optimized models |
+| Planning | task(subagent_type="plan", load_skills=[], run_in_background=false) | Parallel task graph + structured TODO list |
+| Hard problem (conventional) | task(subagent_type="oracle", load_skills=[], run_in_background=false) | Architecture, debugging, complex logic |
+| Hard problem (non-conventional) | task(category="artistry", load_skills=[...], run_in_background=true) | Different approach needed |
+| Implementation | task(category="...", load_skills=[...], run_in_background=true) | Domain-optimized models |
 
 **CATEGORY + SKILL DELEGATION:**
 \`\`\`
 // Frontend work
-task(category="visual-engineering", load_skills=["frontend-ui-ux"])
+task(category="visual-engineering", load_skills=["frontend-ui-ux"], run_in_background=true)
 
 // Complex logic
-task(category="ultrabrain", load_skills=["typescript-programmer"])
+task(category="ultrabrain", load_skills=["typescript-programmer"], run_in_background=true)
 
 // Quick fixes
-task(category="quick", load_skills=["git-master"])
+task(category="quick", load_skills=["git-master"], run_in_background=true)
 \`\`\`
 
 **YOU SHOULD ONLY DO IT YOURSELF WHEN:**

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

@@ -156,7 +156,7 @@ TELL THE USER WHAT AGENTS YOU WILL LEVERAGE NOW TO SATISFY USER'S REQUEST.
 | Architecture decision needed | MUST call plan agent |
 
 \`\`\`
-task(subagent_type="plan", load_skills=[], prompt="<gathered context + user request>")
+task(subagent_type="plan", load_skills=[], run_in_background=false, prompt="<gathered context + user request>")
 \`\`\`
 
 ### SESSION CONTINUITY WITH PLAN AGENT (CRITICAL)
@@ -165,9 +165,9 @@ task(subagent_type="plan", load_skills=[], prompt="<gathered context + user requ
 
 | Scenario | Action |
 |----------|--------|
-| Plan agent asks clarifying questions | \`task(session_id="{returned_session_id}", load_skills=[], prompt="<your answer>")\` |
-| Need to refine the plan | \`task(session_id="{returned_session_id}", load_skills=[], prompt="Please adjust: <feedback>")\` |
-| Plan needs more detail | \`task(session_id="{returned_session_id}", load_skills=[], prompt="Add more detail to Task N")\` |
+| 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")\` |
 
 **FAILURE TO CALL PLAN AGENT = INCOMPLETE WORK.**
 
@@ -183,10 +183,10 @@ task(subagent_type="plan", load_skills=[], prompt="<gathered context + user requ
 |-----------|--------|-----|
 | Codebase exploration | task(subagent_type="explore", load_skills=[], run_in_background=true) | Parallel, context-efficient |
 | Documentation lookup | task(subagent_type="librarian", load_skills=[], run_in_background=true) | Specialized knowledge |
-| Planning | task(subagent_type="plan", load_skills=[]) | Parallel task graph + structured TODO list |
-| Hard problem (conventional) | task(subagent_type="oracle", load_skills=[]) | Architecture, debugging, complex logic |
-| Hard problem (non-conventional) | task(category="artistry", load_skills=[...]) | Different approach needed |
-| Implementation | task(category="...", load_skills=[...]) | Domain-optimized models |
+| Planning | task(subagent_type="plan", load_skills=[], run_in_background=false) | Parallel task graph + structured TODO list |
+| Hard problem (conventional) | task(subagent_type="oracle", load_skills=[], run_in_background=false) | Architecture, debugging, complex logic |
+| Hard problem (non-conventional) | task(category="artistry", load_skills=[...], run_in_background=true) | Different approach needed |
+| Implementation | task(category="...", load_skills=[...], run_in_background=true) | Domain-optimized models |
 
 **YOU SHOULD ONLY DO IT YOURSELF WHEN:**
 - Task is trivially simple (1-2 lines, obvious change)

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

@@ -71,9 +71,9 @@ Use these when they provide clear value based on the decision framework above:
 |----------|-------------|------------|
 | explore agent | Need codebase patterns you don't have | \`task(subagent_type="explore", load_skills=[], run_in_background=true, ...)\` |
 | librarian agent | External library docs, OSS examples | \`task(subagent_type="librarian", load_skills=[], run_in_background=true, ...)\` |
-| oracle agent | Stuck on architecture/debugging after 2+ attempts | \`task(subagent_type="oracle", load_skills=[], ...)\` |
-| plan agent | Complex multi-step with dependencies (5+ steps) | \`task(subagent_type="plan", load_skills=[], ...)\` |
-| task category | Specialized work matching a category | \`task(category="...", load_skills=[...])\` |
+| oracle agent | Stuck on architecture/debugging after 2+ attempts | \`task(subagent_type="oracle", load_skills=[], run_in_background=false, ...)\` |
+| plan agent | Complex multi-step with dependencies (5+ steps) | \`task(subagent_type="plan", load_skills=[], run_in_background=false, ...)\` |
+| task category | Specialized work matching a category | \`task(category="...", load_skills=[...], run_in_background=true)\` |
 
 <tool_usage_rules>
 - Prefer tools over internal knowledge for fresh or user-specific data

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

@@ -117,7 +117,7 @@ Each TODO item MUST include:
 
 | Wave | Tasks | Dispatch Command |
 |------|-------|------------------|
-| 1 | 1, 4 | \`task(category="...", load_skills=[...], run_in_background=false)\` × 2 |
+| 1 | 1, 4 | \`task(category="...", load_skills=[...], run_in_background=true)\` × 2 |
 | 2 | 2, 3, 5 | \`task(...)\` × 3 after Wave 1 completes |
 | 3 | 6 | \`task(...)\` final integration |
 

src/hooks/ralph-loop/oracle-verification-detector.ts

@@ -1,3 +1,4 @@
+import { stripInvisibleAgentCharacters } from "../../shared/agent-display-names"
 import { ULTRAWORK_VERIFICATION_PROMISE } from "./constants"
 
 export interface OracleVerificationEvidence {
@@ -54,7 +55,7 @@ export function isOracleVerified(text: string): boolean {
 		return false
 	}
 
-	const isOracleAgent = evidence.agent.toLowerCase() === "oracle"
+	const isOracleAgent = stripInvisibleAgentCharacters(evidence.agent).toLowerCase() === "oracle"
 	const isVerifiedPromise = evidence.promise === ULTRAWORK_VERIFICATION_PROMISE
 
 	return isOracleAgent && isVerifiedPromise
@@ -62,7 +63,7 @@ export function isOracleVerified(text: string): boolean {
 
 export function extractOracleSessionID(text: string): string | undefined {
 	const evidence = parseOracleVerificationEvidence(text)
-	if (!evidence || evidence.agent.toLowerCase() !== "oracle") {
+	if (!evidence || stripInvisibleAgentCharacters(evidence.agent).toLowerCase() !== "oracle") {
 		return undefined
 	}
 

src/plugin/tool-execute-after.ts

@@ -1,6 +1,7 @@
 import { consumeToolMetadata } from "../features/tool-metadata-store"
 import type { CreatedHooks } from "../create-hooks"
 import { log } from "../shared"
+import { stripInvisibleAgentCharacters } from "../shared/agent-display-names"
 import type { PluginContext } from "./types"
 import { readState, writeState } from "../hooks/ralph-loop/storage"
 
@@ -60,7 +61,7 @@ export function createToolExecuteAfterHandler(args: {
       const verificationAttemptId = prompt?.match(VERIFICATION_ATTEMPT_PATTERN)?.[1]?.trim()
       const loopState = directory ? readState(directory) : null
       const isVerificationContext =
-        agent === "oracle"
+        (agent ? stripInvisibleAgentCharacters(agent) : agent) === "oracle"
         && !!sessionId
         && !!directory
         && loopState?.active === true

src/plugin/tool-execute-before.ts

@@ -4,6 +4,7 @@ import { randomUUID } from "node:crypto"
 import { getMainSessionID } from "../features/claude-code-session-state"
 import { clearBoulderState } from "../features/boulder-state"
 import { log } from "../shared"
+import { stripInvisibleAgentCharacters } from "../shared/agent-display-names"
 import { resolveSessionAgent } from "./session-agent-resolver"
 import { parseRalphLoopArguments } from "../hooks/ralph-loop/command-arguments"
 import { ULTRAWORK_VERIFICATION_PROMISE } from "../hooks/ralph-loop/constants"
@@ -109,7 +110,7 @@ export function createToolExecuteBeforeHandler(args: {
       }
 
       const normalizedSubagentType =
-        typeof argsObject.subagent_type === "string" ? argsObject.subagent_type : undefined
+        typeof argsObject.subagent_type === "string" ? stripInvisibleAgentCharacters(argsObject.subagent_type) : undefined
       const prompt = typeof argsObject.prompt === "string" ? argsObject.prompt : ""
       const loopState = typeof ctx.directory === "string" ? readState(ctx.directory) : null
       const shouldInjectOracleVerification =

src/shared/agent-tool-restrictions.ts

@@ -1,3 +1,5 @@
+import { stripInvisibleAgentCharacters } from "./agent-display-names"
+
 /**
  * Agent tool restrictions for session.prompt calls.
  * OpenCode SDK's session.prompt `tools` parameter expects boolean values.
@@ -45,13 +47,15 @@ const AGENT_RESTRICTIONS: Record<string, Record<string, boolean>> = {
 }
 
 export function getAgentToolRestrictions(agentName: string): Record<string, boolean> {
-  return AGENT_RESTRICTIONS[agentName]
-    ?? Object.entries(AGENT_RESTRICTIONS).find(([key]) => key.toLowerCase() === agentName.toLowerCase())?.[1]
+  const stripped = stripInvisibleAgentCharacters(agentName)
+  return AGENT_RESTRICTIONS[stripped]
+    ?? Object.entries(AGENT_RESTRICTIONS).find(([key]) => key.toLowerCase() === stripped.toLowerCase())?.[1]
     ?? {}
 }
 
 export function hasAgentToolRestrictions(agentName: string): boolean {
-  const restrictions = AGENT_RESTRICTIONS[agentName]
-    ?? Object.entries(AGENT_RESTRICTIONS).find(([key]) => key.toLowerCase() === agentName.toLowerCase())?.[1]
+  const stripped = stripInvisibleAgentCharacters(agentName)
+  const restrictions = AGENT_RESTRICTIONS[stripped]
+    ?? Object.entries(AGENT_RESTRICTIONS).find(([key]) => key.toLowerCase() === stripped.toLowerCase())?.[1]
   return restrictions !== undefined && Object.keys(restrictions).length > 0
 }

src/shared/agent-variant.ts

@@ -1,4 +1,5 @@
 import type { OhMyOpenCodeConfig } from "../config"
+import { stripInvisibleAgentCharacters } from "./agent-display-names"
 import { AGENT_MODEL_REQUIREMENTS, CATEGORY_MODEL_REQUIREMENTS } from "./model-requirements"
 
 export function resolveAgentVariant(
@@ -9,12 +10,13 @@ export function resolveAgentVariant(
     return undefined
   }
 
+  const stripped = stripInvisibleAgentCharacters(agentName)
   const agentOverrides = config.agents as
     | Record<string, { variant?: string; category?: string }>
     | undefined
   const agentOverride = agentOverrides
-    ? agentOverrides[agentName]
-      ?? Object.entries(agentOverrides).find(([key]) => key.toLowerCase() === agentName.toLowerCase())?.[1]
+    ? agentOverrides[stripped]
+      ?? Object.entries(agentOverrides).find(([key]) => key.toLowerCase() === stripped.toLowerCase())?.[1]
     : undefined
   if (!agentOverride) {
     return undefined
@@ -37,18 +39,19 @@ export function resolveVariantForModel(
   agentName: string,
   currentModel: { providerID: string; modelID: string },
 ): string | undefined {
+  const stripped = stripInvisibleAgentCharacters(agentName)
   const agentOverrides = config.agents as
     | Record<string, { variant?: string; category?: string }>
     | undefined
   const agentOverride = agentOverrides
-    ? agentOverrides[agentName]
-      ?? Object.entries(agentOverrides).find(([key]) => key.toLowerCase() === agentName.toLowerCase())?.[1]
+    ? agentOverrides[stripped]
+      ?? Object.entries(agentOverrides).find(([key]) => key.toLowerCase() === stripped.toLowerCase())?.[1]
     : undefined
   if (agentOverride?.variant) {
     return agentOverride.variant
   }
 
-  const agentRequirement = AGENT_MODEL_REQUIREMENTS[agentName]
+  const agentRequirement = AGENT_MODEL_REQUIREMENTS[stripped]
   if (agentRequirement) {
     return findVariantInChain(agentRequirement.fallbackChain, currentModel)
   }

src/tools/call-omo-agent/tools.ts

@@ -6,7 +6,7 @@ import type { CategoriesConfig, AgentOverrides } from "../../config/schema"
 import type { DelegatedModelConfig } from "../../shared/model-resolution-types"
 import type { FallbackEntry } from "../../shared/model-requirements"
 import { AGENT_MODEL_REQUIREMENTS } from "../../shared/model-requirements"
-import { getAgentConfigKey } from "../../shared/agent-display-names"
+import { getAgentConfigKey, stripInvisibleAgentCharacters } from "../../shared/agent-display-names"
 import { normalizeFallbackModels } from "../../shared/model-resolver"
 import { buildFallbackChainFromModels } from "../../shared/fallback-chain-from-models"
 import { log } from "../../shared"
@@ -104,20 +104,21 @@ export function createCallOmoAgent(
       const toolCtx = toolContext as ToolContextWithMetadata
       log(`[call_omo_agent] Starting with agent: ${args.subagent_type}, background: ${args.run_in_background}`)
 
-      // Case-insensitive agent validation - allows "Explore", "EXPLORE", "explore" etc.
+      // Strip ZWSP and case-insensitive agent validation - allows "Explore", "EXPLORE", "explore" etc.
+      const strippedAgentType = stripInvisibleAgentCharacters(args.subagent_type)
       if (
         !ALLOWED_AGENTS.some(
-          (name) => name.toLowerCase() === args.subagent_type.toLowerCase(),
+          (name) => name.toLowerCase() === strippedAgentType.toLowerCase(),
         )
       ) {
         return `Error: Invalid agent type "${args.subagent_type}". Only ${ALLOWED_AGENTS.join(", ")} are allowed.`
       }
 
-      const normalizedAgent = args.subagent_type.toLowerCase() as AllowedAgentType
+      const normalizedAgent = strippedAgentType.toLowerCase() as AllowedAgentType
       args = { ...args, subagent_type: normalizedAgent }
 
       // Check if agent is disabled
-      if (disabledAgents.some((disabled) => disabled.toLowerCase() === normalizedAgent)) {
+      if (disabledAgents.some((disabled) => stripInvisibleAgentCharacters(disabled).toLowerCase() === normalizedAgent)) {
         return `Error: Agent "${normalizedAgent}" is disabled via disabled_agents configuration. Remove it from disabled_agents in your ${CONFIG_BASENAME}.json to use it.`
       }
 

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

@@ -8,6 +8,7 @@ import {
 } from "../../shared/model-suggestion-retry"
 import { formatDetailedError } from "./error-formatting"
 import { getAgentToolRestrictions } from "../../shared/agent-tool-restrictions"
+import { stripInvisibleAgentCharacters } from "../../shared/agent-display-names"
 import { applySessionPromptParams } from "../../shared/session-prompt-params-helpers"
 import { setSessionTools } from "../../shared/session-tools-store"
 import { createInternalAgentTextPart } from "../../shared/internal-initiator-marker"
@@ -41,7 +42,7 @@ function buildPromptGenerationParams(model: DelegatedModelConfig | undefined): R
 }
 
 function isOracleAgent(agentToUse: string): boolean {
-  return agentToUse.toLowerCase() === "oracle"
+  return stripInvisibleAgentCharacters(agentToUse).toLowerCase() === "oracle"
 }
 
 function isUnexpectedEofError(error: unknown): boolean {
@@ -80,7 +81,7 @@ export async function sendSyncPrompt(
   const promptArgs = {
     path: { id: input.sessionID },
     body: {
-      agent: input.agentToUse.replace(/^\u200B+/, ""),
+      agent: stripInvisibleAgentCharacters(input.agentToUse),
       system: input.systemContent,
       tools,
       parts: [createInternalAgentTextPart(effectivePrompt)],