[FEATURE] Allow rough effort estimates for software projects

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

Claude Code is explicitly configured to refuse any time or effort estimates for software tasks. When asked for approximate effort assessments, it responds with a blanket refusal, even when the user
explicitly acknowledges the inherent uncertainty.

This is problematic because:

1. Core use case of a coding assistant: Claude Code assists with software development. Project planning, including effort estimation, is an integral part of software development. Refusing to provide any

guidance here limits the tool's usefulness significantly.

1. Users understand AI limitations: Users working with AI tools are generally aware that outputs are approximations and require validation. This applies equally to code suggestions, architectural decisions,

and effort estimates. There's no reason to single out estimates for special treatment.

1. Comparative disadvantage: Any experienced developer can provide rough estimates. Claude Code has access to the actual codebase and could potentially provide more informed estimates than a developer giving

a gut feeling.

1. Practical impact: Users need rough indicators for personal time planning, evaluating contractor proposals, prioritizing tasks, and communicating with stakeholders.

Proposed Solution

Allow Claude Code to provide effort estimates with appropriate caveats, similar to how it handles other uncertain outputs. For example:

"Based on the scope I've analyzed, this appears to be a moderate effort involving X files and Y components. Experienced developers typically estimate similar tasks at [range], but actual effort depends
heavily on familiarity with the codebase, testing requirements, and unforeseen complications."

The key is to provide useful guidance while being transparent about uncertainty - the same approach Claude takes with all other outputs.

Alternative Solutions

Currently, users must ask Claude Code to analyze the codebase, then take that analysis to another source for estimation. This creates unnecessary friction and reduces productivity.

Priority

High - Significant impact on productivity

Feature Category

API and model interactions

Use Case Example

_No response_

Additional Context

_No response_