Subagents don't inherit bypassPermissions mode from parent session

[sorfanit]

Description

When the parent session is configured with "defaultMode": "bypassPermissions" in ~/.claude/settings.json, subagents spawned via the Agent tool still prompt for tool permissions. This defeats the purpose of bypass mode and creates a disruptive experience — every subagent invocation can trigger permission prompts that the user has already opted out of.

Steps to Reproduce

1. Set "defaultMode": "bypassPermissions" in ~/.claude/settings.json
2. Also set "skipDangerousModePermissionPrompt": true
3. Start a Claude Code session — main agent correctly bypasses all permission prompts
4. Invoke a task that uses the Agent tool (subagent)
5. The subagent prompts for permissions on tools the parent session would silently allow

Expected Behavior

Subagents should inherit the parent session's permission mode, including bypassPermissions and all allow patterns.

Actual Behavior

Subagents prompt for permissions as if running in a more restrictive mode.

Environment

• Claude Code version: 2.1.81
• OS: Windows 11 Pro (10.0.26200)
• Shell: bash (Git Bash)

Notes

Persistent issue across multiple sessions over several days. Adding broad allow patterns (e.g., Bash(*)) as a workaround doesn't fully resolve it since bypassPermissions should already cover everything.

[github-actions]

Found 3 possible duplicate issues:

1. --dangerously-skip-permissions does not propagate to context: fork sub-agents #26083
2. [BUG] Agent Teams teammates ignore bypassPermissions for Bash and don't inherit project settings.local.json #26479
3. Agent teams: bypassPermissions mode revokes Bash access mid-session #36225

This issue will be automatically closed as a duplicate in 3 days.

• If your issue is a duplicate, please close it and 👍 the existing issue instead
• To prevent auto-closure, add a comment or 👎 this comment

🤖 Generated with Claude Code

[yurukusa]

The Agent tool has a mode parameter that can explicitly set the permission mode for spawned subagents. Until the inheritance bug is fixed, you can work around this by instructing Claude to use mode: "bypassPermissions" when spawning agents.
In your CLAUDE.md, add:

When using the Agent tool, always set mode: "bypassPermissions"

This tells Claude to pass the mode explicitly on every Agent invocation, bypassing the broken inheritance.
If you want a more enforceable approach, a PreToolUse hook on Agent can inject the mode:

INPUT=$(cat)
TOOL=$(echo "$INPUT" | jq -r '.tool_name // empty' 2>/dev/null)
[[ "$TOOL" != "Agent" ]] && exit 0
MODE=$(echo "$INPUT" | jq -r '.tool_input.mode // empty' 2>/dev/null)
if [[ -z "$MODE" || "$MODE" == "null" ]]; then
    echo "NOTE: Agent spawned without explicit mode parameter." >&2
    echo "Add mode: \"bypassPermissions\" to the Agent call." >&2
fi
exit 0

This won't block the call, but will warn when agents are spawned without an explicit mode, helping you catch the inconsistency.