[FEATURE] Human-readable IDs for background tool tasks

Preflight Checklist

• [x] I have searched existing requests and this feature hasn't been requested yet
• [x] This is a single feature request (not multiple features)

Problem Statement

When invoking Bash (or other tools) with run_in_background: true, the returned task ID is a short opaque random string (e.g. bclalsipw, bljju2ana). With more than one background task in flight, it's hard for both the agent and the user to remember which is which — every reference needs a sidecar phrase like "the nuget polling one" or "the workflow watcher."

Proposed Solution

Could the runtime either:

1. Derive the ID from the task's description (slugify, dedupe with a counter): nuget-indexing-poll-1, workflow-run-monitor-2. The opaque ID can stay as the canonical handle underneath; the slug is just a display alias the

agent and user can reference.

1. Accept an optional name/id parameter on Bash/Agent so the agent can pick a meaningful handle when it starts the task.

Either would make multi-task sessions easier to navigate — particularly when the user asks "what is <id>?", which has happened in practice.

Alternative Solutions

_No response_

Priority

Medium - Would be very helpful

Feature Category

CLI commands and flags

Use Case Example

Example Scenario:

1. Task Claude Code with publishing packages and monitoring for success
2. Task Claude Code with running static analysis tools
3. Task Claude Code with running the test suite and monitoring

These may all be long running bash processes that CC will spawn. Each would be given a random string process ID that CC will use for discussion. The user must ask what each is if/when they are referenced later in conversation. If they could be auto-generated or passed by parameter, the ambiguity could be avoided.

Additional Context

_No response_