Issue Station
Sign In
Tools hereToolbox

claude-code - 💡(How to fix) Fix [BUG] Subagents look up agent dependencies as skills instead of using the Agent tool [2 comments, 2 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
anthropics/claude-code#55086•Fetched 2026-05-01 05:46:36
View on GitHub
Comments
2
Participants
2
Timeline
7
Reactions
0
Author
D029009D029009
Participants
D029009D029009github-actions[bot]github-actions[bot]
Timeline (top)
labeled ×5commented ×2

Error Message

Since approximately 00:00 CET on 2026-04-29, subagents that previously invoked other agents via the Agent tool now check the skills list (from the system reminder) instead. When the target agent is not registered as a skill, the subagent stops with an error.

Error Messages/Logs

Fix Action

Fix / Workaround

The agent definitions are unchanged from when they worked The skills system reminder (listing available skills) appears to be what causes the confusion — subagents now treat it as the authoritative list of callable agents Workaround: none found

Code Example

The orchestrator reports: "The skill xyz is not available in the current session" and stops. Even when the instructions explicitly say "use the Agent tool with subagent_type worker", the model still attempts a skill lookup instead.
RAW_BUFFERClick to expand / collapse

Preflight Checklist

  • I have searched existing issues and this hasn't been reported yet
  • This is a single bug report (please file separate reports for different bugs)
  • I am using the latest version of Claude Code

What's Wrong?

Since approximately 00:00 CET on 2026-04-29, subagents that previously invoked other agents via the Agent tool now check the skills list (from the system reminder) instead. When the target agent is not registered as a skill, the subagent stops with an error.

The agent definitions are unchanged from when they worked The skills system reminder (listing available skills) appears to be what causes the confusion — subagents now treat it as the authoritative list of callable agents Workaround: none found

What Should Happen?

The orchestrator invokes subagent via the Agent tool, giving it an isolated context, as it did before.

Error Messages/Logs

The orchestrator reports: "The skill xyz is not available in the current session" and stops. Even when the instructions explicitly say "use the Agent tool with subagent_type worker", the model still attempts a skill lookup instead.

Steps to Reproduce

  1. Create ~/.claude/agents/worker.md — a simple agent that does some work
  2. Create ~/.claude/agents/orchestrator.md — an agent whose instructions say "call subagent worker for each item"
  3. Launch the orchestrator as a subagent via the Agent tool
  4. Observe: the orchestrator checks the skills list for worker, fails to find it, and stops instead of using the Agent tool

Claude Model

Sonnet (default)

Is this a regression?

Yes, this worked in a previous version

Last Working Version

day before yesterday

Claude Code Version

 2.1.123 (Claude Code)

Platform

Other

Operating System

Windows

Terminal/Shell

Windows Terminal

Additional Information

No response

extent analysis

TL;DR

The issue can be mitigated by ensuring that subagents are properly registered as skills or by modifying the orchestrator to bypass the skills list check.

Guidance

  • Review the agent definitions and skills list to ensure that subagents are correctly registered and configured.
  • Verify that the worker.md agent is correctly defined and located in the ~/.claude/agents/ directory.
  • Check the orchestrator's instructions to ensure that it is correctly calling the subagent using the Agent tool.
  • Consider modifying the orchestrator to explicitly specify the subagent type and bypass the skills list check.

Example

No code snippet is provided as the issue does not contain sufficient information to create a specific example.

Notes

The issue appears to be a regression, and the last working version was the day before yesterday. The cause of the issue is likely related to a change in how the Agent tool interacts with the skills system.

Recommendation

Apply workaround: Modify the orchestrator to explicitly specify the subagent type and bypass the skills list check, as this is a more targeted solution that can be implemented immediately.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#inference speed#output truncation#response parsing#generation error#database connection

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

claude-code - 💡(How to fix) Fix [BUG] Subagents look up agent dependencies as skills instead of using the Agent tool [2 comments, 2 participants]