How to use the Context Builder for Feature Research agent skill

What an agent skill is and what Context Builder makes the agent do

An agent skill is a folder containing a SKILL.md file that gives an AI agent specific instructions for one kind of task. When the skill is installed, the agent can follow those instructions when a request matches its description.

Context Builder's job is to gather and distill background material for a feature or initiative. It produces two kinds of output: a feature-specific background.md in the feature folder, and updates to shared context documents when it finds synthesized knowledge that is useful across several features, such as a customer requirements summary or a competitive analysis.

• Asks the product manager what feature is being researched, which aspects matter most, and any specific meetings, customers or competitors to focus on.
• Checks for an existing validation.md in the feature folder, then shared context, meetings, competitor research, regulatory sources, customer requests and existing specs.
• Writes background.md with sections for customer drivers, competitive landscape, regulatory considerations, persona context and internal context.
• Flags contradictions and dates its sources.

Install Context Builder

Download the ZIP from this page and unzip it. The skill is a folder that contains SKILL.md. Place that folder in one of these locations, depending on how you want it available:

For a personal install across projects, unzip it into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder.

If you use the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.

• Personal (all projects): ~/.claude/skills/
• Single project: .claude/skills/ inside the project folder
• Claude apps: Settings → Capabilities → Skills, then upload

Set up your source folders

Context Builder searches the files you already have. The source describes a layered structure: raw sources in reference/ (meetings, PDFs, documents), synthesized markdown shared across features in context/, and feature-specific background.md in each feature folder.

The skill works best when these folders exist and your meeting notes, competitor research and customer requests are stored in them. It is not a replacement for organizing that material. If the folders are empty or disorganized, expect thin output.

• reference/ for raw meetings, PDFs and documents
• context/ for synthesized shared context docs
• A feature folder for each initiative, optionally with validation.md

Trigger the skill with a natural request

You do not need a special command. Describe the feature and ask for background. The agent will ask follow-up questions about the feature, focus areas and specific sources before searching.

Name the feature folder and any customers or competitors you care about. This narrows the search and reduces irrelevant material.

• "Build background context for the bulk export feature."
• "Gather competitor and customer research for our new onboarding initiative and write background.md."
• "Pull together regulatory and customer context for the invoicing reminders feature, focusing on the EU customers mentioned in our meeting notes."

Review what the skill produces

The background.md file uses front matter that lists the meeting files, competitive references, regulatory sources and customer paths it used. Each section should point back to those sources. Check that the links resolve and that the sources are recent enough for the topic, since the skill is instructed to check roughly the last 3 to 6 months of meetings and to date its sources.

Look at the flagged contradictions and the 'surprising or non-obvious' points first. These are where a human reviewer adds the most value.

• Confirm each claim links to a real file.
• Check source dates against the topic.
• Read any contradictions the skill flagged before making decisions.

Safety notes: scripts and network access

According to the listing, this skill runs no scripts and uses no network access. It works by reading and writing files in your project folders. Because it can write shared context docs, review those edits before committing them, especially if other features depend on them.

The skill does not give legal or regulatory compliance advice. It summarizes what is in the files you provide, so any regulatory content should be checked by a qualified person.

• No scripts are run and no network use is listed.
• Shared context edits affect every feature that references them; review them.
• Regulatory summaries are not compliance advice.