fix(thread): 🐛 hide custom agent tool calls and display proper names by jorben · Pull Request #204 · tiylabs/tiycode

jorben · 2026-05-25T02:54:01Z

Summary

Fix custom agent agent_{slug} tool calls leaking into the timeline alongside helper cards
Display user-defined custom agent names instead of raw helper_custom_{slug} strings
Unify isRuntimeOrchestrationToolName to use startsWith("agent_") pattern, eliminating duplicate implementations

Test Plan

npm run typecheck passes
npm run test:unit passes (752 tests)
Create a custom agent (e.g. slug=refactor, name="Refactor Agent"), trigger a call, verify timeline no longer shows a duplicate agent_refactor tool card
Verify the helper card displays the custom agent's user-defined name

🤖 Generated with TiyCode

Unify isRuntimeOrchestrationToolName to use startsWith("agent_") pattern so custom agent_{slug} calls are hidden from timeline alongside built-in agent_explore/agent_review. Add slugToNameMap support to formatHelperKind so custom agents show user-defined names instead of raw helper_custom_{slug}.

github-actions · 2026-05-25T02:56:32Z

AI Code Review Summary

PR: #204 (fix(thread): 🐛 hide custom agent tool calls and display proper names)
Preferred language: English

Overall Assessment

Detected 3 actionable findings, prioritize CRITICAL/HIGH before merge.

Major Findings by Severity

MEDIUM (1)
- src/shared/constants/tool-names.ts:33 - Prefix-based identification of privileged orchestration tools
LOW (2)
- src/modules/workbench-shell/ui/runtime-thread-surface-helpers.ts:118 - Fallback handling for empty slug in custom helper kind
- src/shared/constants/tool-names.ts:32 - Namespace Collision Risk with Prefix-Based Tool Classification

Actionable Suggestions

Buddy, consider fallback checks like return customName ?? (slug ? capitalizeSlug(slug) : 'Custom Agent'); on line 121 in runtime-thread-surface-helpers.ts to guarantee a visible fallback name.
Buddy, consider adding a registration check/validation that blocks standard (non-orchestration) custom tools from using the reserved 'agent_' prefix.
Verify where isRuntimeOrchestrationToolName is called in both the frontend and backend. Ensure it is not used as a security/consent-bypass check for third-party or custom user tools.
Implement a robust, strictly controlled registry for orchestration-level tools rather than relying on string prefixes.

Potential Risks

Empty rendering labels on custom subagent elements if a malformed prefix without a slug is received.
Third-party or user-configured tools that happen to use the 'agent_' prefix will be treated as runtime orchestration tools, which may cause execution routing failures.
Malicious extensions or custom tools bypassing security boundaries or confirmation dialogs if those boundaries rely on the isRuntimeOrchestrationTool classification.

Test Suggestions

Run npm run test:unit to verify the vitest suite handles custom helper formatting behaviors.
Ensure that backend integration tests verify how the agent session runner handles unknown tools prefixed with 'agent_' to confirm failure modes are graceful.
Add integration tests in the Tauri/Rust core runtime backend to ensure tool name spoofing (e.g. prefixing with 'agent_') cannot bypass security, capability checks, or prompt consent checks.

File-Level Coverage Notes

src/modules/workbench-shell/ui/runtime-thread-surface-helpers.ts: The changes successfully add support for custom subagent helpers by introducing custom slug-to-name mapping and automated fallback capitalization. (Overall robust functional code.)
src/modules/workbench-shell/ui/runtime-thread-surface-helpers.test.ts: Excellent testing suite that achieves robust coverage of standard cases, unknown helpers, dynamic map fallbacks, and edge cases. (Great use of a mock helper generator function.)
src/modules/workbench-shell/ui/runtime-thread-surface-state.ts: A clean refactoring to centralize orchestration tool identifier matching into the shared constants folder. (Enhances code consistency and maintainability.)
src/modules/workbench-shell/ui/runtime-thread-surface.tsx: Efficient mapping of custom subagent slugs using React useMemo to prevent unnecessary re-computations during render loop iterations. (Correctly integrates the new slugToNameMap argument across helper formatting calls.)
src/shared/constants/tool-names.ts: Refactoring tool verification from an allowlist to a prefix-based match introduces a potential security risk of authorization bypass if orchestration tools possess elevated capabilities. (Buddy, we should ensure the backend enforcement does not solely rely on this frontend helper for determining tool security boundaries.)
src/shared/constants/tool-names.test.ts: The test changes align with the implementation change but confirm that arbitrary extensions (e.g., 'agent_refactor') are now classified as orchestration tools. (The tests verify the prefix matching behavior but remove previous negative test cases for extended matches, validating that prefix matching is indeed the new intended behavior.)

Inline Downgraded Items (processed but not inline)

None

Coverage Status

Target files: 6
Covered files: 6
Uncovered files: 0
No-patch/binary covered as file-level: 0
Findings with unknown confidence (N/A): 0

Uncovered list:

None

No-patch covered list:

None

Runtime/Budget

Rounds used: 1/4
Planned batches: 2
Executed batches: 2
Sub-agent runs: 3
Planner calls: 1
Reviewer calls: 4
Model calls: 5/64
Structured-output summary-only degradation: NO

github-actions

Automated PR review completed.

Findings kept: 4
Findings with unknown confidence: 0
Inline comments attempted: 4
Target files: 5
Covered files: 5
Uncovered files: 0
See the summary comment for detailed analysis and coverage details.

Sign in to view

+    .join(" ");
+}
+
+export function formatHelperKind(


Sign in to view

  const agentProfiles = useStore(settingsStore, (s) => s.agentProfiles);
  const customSubagents = useStore(settingsStore, (s) => s.customSubagents, shallowEqual);
+  const customAgentSlugToName = useMemo(
+    () => new Map(customSubagents.map((a) => [a.slug, a.name])),


Sign in to view


 export function isRuntimeOrchestrationToolName(toolName: string): boolean {
-  return (RUNTIME_ORCHESTRATION_TOOLS as ReadonlyArray<string>).includes(toolName);
+  return toolName.startsWith(RUNTIME_ORCHESTRATION_TOOL_PREFIX);


Sign in to view


 export function isRuntimeOrchestrationToolName(toolName: string): boolean {
-  return (RUNTIME_ORCHESTRATION_TOOLS as ReadonlyArray<string>).includes(toolName);
+  return toolName.startsWith(RUNTIME_ORCHESTRATION_TOOL_PREFIX);


…tests Require at least one character after the agent_ prefix so bare "agent_" is no longer classified as a runtime orchestration tool. Add unit tests for formatHelperKind, formatHelperName, formatHelperSummary covering built-in kinds, custom agent name lookup, slug fallback capitalization, and edge cases.

github-actions

Automated PR review completed.

Findings kept: 2
Findings with unknown confidence: 0
Inline comments attempted: 2
Target files: 6
Covered files: 6
Uncovered files: 0
See the summary comment for detailed analysis and coverage details.

Sign in to view

@@ -26,7 +29,10 @@ export const DEFAULT_COLLAPSED_TOOLS = [
 ] as const;

 export function isRuntimeOrchestrationToolName(toolName: string): boolean {


Sign in to view

-export function formatHelperKind(kind: string) {
+const CUSTOM_PREFIX = "helper_custom_";
+
+function capitalizeSlug(slug: string) {


Add .filter(Boolean) after split("-") so that hypothetical consecutive hyphens or trailing hyphens do not produce stray empty-string segments.

github-actions

Automated PR review completed.

Findings kept: 3
Findings with unknown confidence: 0
Inline comments attempted: 3
Target files: 6
Covered files: 6
Uncovered files: 0
See the summary comment for detailed analysis and coverage details.

github-actions · 2026-05-25T03:59:27Z

 export function isRuntimeOrchestrationToolName(toolName: string): boolean {
-  return (RUNTIME_ORCHESTRATION_TOOLS as ReadonlyArray<string>).includes(toolName);
+  return (
+    toolName.startsWith(RUNTIME_ORCHESTRATION_TOOL_PREFIX)


[MEDIUM] Prefix-based identification of privileged orchestration tools

Buddy, the function isRuntimeOrchestrationToolName has been changed from checking an explicit allowlist to checking a simple prefix ('agent_'). If runtime orchestration tools receive elevated privileges, bypass consent prompts, or run in a different isolation boundary, any untrusted custom tool or plugin could gain these privileges simply by prefixing its name with 'agent_'.

Suggestion: Instead of relying on a string prefix to identify privileged orchestration tools, maintain a strict and secure registry of authorized orchestration tools, or require explicit registration with verification rather than basic prefix matching.

Risk: Untrusted or malicious extensions/tools can masquerade as orchestration tools, potentially escalating privileges, bypassing validation, or execution policies.

Confidence: 0.85

[From SubAgent: security]

github-actions · 2026-05-25T03:59:27Z

    case "helper_review":
      return "Review Agent";
    default:
+      if (kind.startsWith(CUSTOM_PREFIX)) {


[LOW] Fallback handling for empty slug in custom helper kind

If the helper kind is exactly 'helper_custom_' (with an empty slug), capitalizeSlug returns an empty string, which causes formatHelperKind to return an empty string. This could result in blank names rendering in the UI.

Suggestion: Add a defensive fallback in formatHelperKind to return a default string like 'Custom Agent' or the raw 'kind' string if the sliced slug is falsy.

Risk: Minor visual glitch (empty label) in the UI if an empty custom helper kind ever reaches the client.

Confidence: 0.90

[From SubAgent: general]

github-actions · 2026-05-25T03:59:27Z


 export function isRuntimeOrchestrationToolName(toolName: string): boolean {
-  return (RUNTIME_ORCHESTRATION_TOOLS as ReadonlyArray<string>).includes(toolName);
+  return (


[LOW] Namespace Collision Risk with Prefix-Based Tool Classification

Switching from an explicit allowlist to a prefix-based check ('agent_') means any tool (including user-defined or third-party extensions) that happens to begin with 'agent_' will be treated as a runtime orchestration tool. This could lead to runtime errors or security/routing misbehaviors if standard tools are mistakenly processed as orchestration tools.

Suggestion: Buddy, make sure there is documentation or validation at the tool registration layer that reserves the 'agent_' prefix exclusively for runtime orchestration tools, or enforce this via system-level validation.

Risk: Custom tools from extensions or integrations prefixed with 'agent_' might be mistakenly routed as runtime orchestration tools, leading to unexpected behavior or execution failures.

Confidence: 0.90

[From SubAgent: general]

github-actions Bot reviewed May 25, 2026

View reviewed changes

fix(thread): 🐛 harden capitalizeSlug against empty segments

9477fe8

Add .filter(Boolean) after split("-") so that hypothetical consecutive hyphens or trailing hyphens do not produce stray empty-string segments.

github-actions Bot reviewed May 25, 2026

View reviewed changes

		@@ -26,7 +29,10 @@ export const DEFAULT_COLLAPSED_TOOLS = [
		] as const;

		export function isRuntimeOrchestrationToolName(toolName: string): boolean {

Conversation

jorben commented May 25, 2026

Summary

Test Plan

Uh oh!

github-actions Bot commented May 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

AI Code Review Summary

Overall Assessment

Major Findings by Severity

Actionable Suggestions

Potential Risks

Test Suggestions

File-Level Coverage Notes

Inline Downgraded Items (processed but not inline)

Coverage Status

Runtime/Budget

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

This comment was marked as outdated.

Uh oh!

This comment was marked as outdated.

Uh oh!

github-actions Bot left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions Bot May 25, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions Bot May 25, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions Bot May 25, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

github-actions Bot commented May 25, 2026 •

edited

Loading