Guide · Agent skill
How to use the Task Effort Estimator agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The estimate skill is an agent skill for Claude Code that produces a structured effort estimate for a task. Agent skills are folders containing a SKILL.md file that teach the agent a specific, repeatable workflow. This guide explains what the estimate skill does, how to install it, how to trigger it with natural requests, and what to check before trusting its numbers.
Task Effort EstimatorDonchitos/Claude-Code-Game-Studios · MIT Download the skill Quick start
- 1Download the ZIP. Download the estimate skill ZIP from this page.
- 2Unzip the skill folder. Unzip it into ~/.claude/skills/ for all projects, or into the project's .claude/skills/ folder for one project.
- 3Or upload in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the skill.
- 4Check your project files. Make sure CLAUDE.md, design/gdd/ and production/sprints/ exist where possible, since they improve the estimate.
- 5Ask for an estimate. Describe the task in plain language, including the systems it touches and any open decisions.
- 6Review the range and confidence. Use the expected figure as the budget and read the confidence explanation and assumptions.
What the estimate skill makes the agent do
When you ask the agent to estimate a task, the skill works through four phases. It first reads your project notes in CLAUDE.md, any related design documents in design/gdd/, and past sprint records in production/sprints/. It then identifies the affected files and measures them by counting non-blank lines and branching statements with Grep. Finally it analyses complexity, scope and risk.
The output is a report with optimistic, expected and pessimistic day estimates, a recommended budget, a confidence level, risk factors, dependencies and a suggested task breakdown. The skill is read-only: it does not change any files.
- Reads CLAUDE.md, design documents and sprint history
- Counts size and branching per affected file
- Produces a day range, recommended budget and confidence level
- Lists risks, dependencies and a sub-task breakdown
- Writes no files
Install the estimate skill from a ZIP or the Claude apps
Download the ZIP for the skill from this page and unzip it. The skill is a folder that contains SKILL.md. To make it available to Claude Code for all your projects, unzip it into ~/.claude/skills/. To make it available only inside one project, unzip it into that project's .claude/skills/ folder.
If you use the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.
- All projects: place the skill folder in ~/.claude/skills/
- One project: place the skill folder in the project's .claude/skills/
- Claude apps: upload under Settings → Capabilities → Skills
Trigger the skill with a natural request
You can ask for an estimate in plain language, such as asking how big a feature is or how many days a refactor would take. The skill takes a task description as its argument, so the more specific the description, the more useful the estimate. If the description is too vague, the agent will ask you for clarification before estimating.
Give the skill the task, the systems it touches and any decisions still open. Open decisions matter because the skill treats an undecided approach as a reason for Low confidence.
- "Estimate the effort to add inventory crafting to the survival system."
- "How big is it to migrate the save file format? Check the design doc first."
- "Give me an estimate with confidence for the new combat UI, based on our sprint history."
Read the estimate report
The report opens with a complexity assessment table, then an effort table with three scenarios. Each scenario states its assumption. The recommended budget is the expected figure, not the optimistic one. Estimates are given in half-day increments.
The confidence level explains which factors drove the result. Check that section and the Notes and Assumptions section before using the numbers. If the report says no sprint history was found, the estimate is not calibrated to your team's velocity.
- Optimistic: everything goes right
- Expected: normal pace, minor issues, one review round (use this as the budget)
- Pessimistic: significant unknowns surface
What the estimate needs and how it measures
The skill requires Claude Code with the Read, Glob and Grep tools. A project with CLAUDE.md, design/gdd/ and production/sprints/ gives the best results, though all three are optional. It runs no scripts and makes no network calls.
The size and branching figures are line and pattern counts, not measured complexity. The branching count includes comments and strings, and the report labels it as an approximation of cyclomatic complexity. Treat these numbers as inputs to judgement, not as measurements of effort.
- Requires: Claude Code with Read, Glob and Grep
- Recommended: CLAUDE.md, design/gdd/, production/sprints/
- Runs scripts: no
- Uses network: no
Safety notes and tips for better estimates
Because the skill is read-only and runs no scripts, it cannot modify your code or sprint records. Its main limitation is accuracy: the output is heuristic. Do not use it as the basis for a precise schedule or a client commitment without your own review.
To get better results, keep sprint records and design documents in the repository, and state open decisions explicitly in your request.
- Keep past sprint records in production/sprints/ so velocity can be used
- Keep design documents in design/gdd/ with clear headings
- Ask for a spike or story split when the report recommends it
- Treat any task over 10 days as a candidate for smaller stories
What people use it for
- Sizing a new gameplay feature before a sprint planning meeting
- Getting a second opinion on how big a refactor is before committing to it
- Identifying risks and a sub-task breakdown for a large feature
- Flagging when a task needs a spike because the approach is undecided
Frequently asked questions
How do I install the estimate agent skill?
Download the ZIP from this page and unzip it. Place the folder in ~/.claude/skills/ to use it in all projects, or in a project's .claude/skills/ folder to use it in one project. In the Claude apps, you can upload it under Settings → Capabilities → Skills.
Is the estimate skill free and what is its licence?
The skill is published under the MIT licence, as listed on this page. The source does not mention pricing, so check the repository for any costs related to Claude Code itself.
What does the estimate skill need to run?
It needs Claude Code with the Read, Glob and Grep tools. A project with CLAUDE.md, design/gdd/ and production/sprints/ gives the best results, but these are optional. It runs no scripts and uses no network.
Does the estimate skill change my files?
No. The skill is read-only and does not write any files. It reads your project notes, design documents and sprint records, then reports its estimate.
Why is my estimate Low confidence?
Low confidence usually means there is no sprint history or documented codebase, or the approach and core requirements are still undecided. The skill treats a range built on an unmade decision as a guess, so resolve open questions and add sprint records to improve it.
How accurate is the estimate compared with a manual estimate?
The output is heuristic and based on line counts and branch counts rather than measured complexity or real velocity data. Use it as a second opinion alongside your team's judgement, and do not use it alone for client commitments.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use the estimate skill if you run a game or software project with CLAUDE.md, design documents and sprint history in the repository, and you want a structured second opinion on task size. Pick another approach if you have no sprint history or need precise schedules or client commitments, since the output is a heuristic range.
Download the skill