Guide · Agent skill
How to use the Idea Lineage Tracer agent skill (idea-lineage)
Updated · written with AI from the project's own documentation and checked against our catalogue data
The idea-lineage skill, also called the Idea Lineage Tracer, lets an AI agent take one idea, phrase or concept page from a gbrain personal knowledge base and reconstruct its history. This guide explains what an agent skill is, how to install this one, how to ask for it in plain language, and what to expect in the answer. By the end you will know how to get a cited account of when an idea appeared, how it changed, and which versions were dropped.
Idea Lineage Tracergarrytan/gbrain · MIT Download the skill Quick start
- 1Download the ZIP. Download the idea-lineage ZIP from this page.
- 2Unzip into a skills folder. Unzip into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder.
- 3Or upload in the Claude apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the skill.
- 4Confirm your brain has the required tools. Make sure your gbrain setup exposes search, query, get_page, list_pages, takes_search, find_contradictions and find_trajectory.
- 5Ask about one idea. Use a natural request such as 'How has my thinking about compounding trust changed?'
- 6Review the cited answer. Check the citations, dates, quotes and confidence labels in each section.
What an agent skill is and what this one makes the agent do
An agent skill is a folder containing a SKILL.md file that tells an AI agent how to perform a specific task. When the skill is installed, the agent can use it when your request matches the skill's triggers.
The idea-lineage skill traces one idea through your brain. It looks for the first mention, the best articulation, the current live version, reversals, contradictions, abandoned branches and related concepts. Every claim must cite a page, date and short quote, and gaps are labeled rather than filled in.
The skill is read-only by default. It saves a lineage page only if you ask for one after seeing the answer.
- Good fit: one idea, phrase or concept page in a long-running gbrain knowledge base.
- Not a fit: broad requests like 'map all my concepts', which route to a different skill (concept-synthesis).
- Not a fit: tracking numeric or status facts over time, such as revenue or job titles, which is handled by find_trajectory.
Install the idea-lineage skill from a ZIP file
Download the ZIP for the idea-lineage skill from this page, then unzip it. You can install it for yourself or for a single project.
For all your projects, unzip it into ~/.claude/skills/. For one project, unzip it into that project's .claude/skills/ folder. Either way, the skill folder should contain SKILL.md.
In the Claude apps you can instead upload the skill under Settings → Capabilities → Skills.
- Personal install: ~/.claude/skills/
- Project install: the project's .claude/skills/ folder
- Claude apps: Settings → Capabilities → Skills
Check that your gbrain setup meets the requirements
The skill needs a gbrain knowledge base that exposes these tools: search, query, get_page, list_pages, takes_search, find_contradictions and find_trajectory. It also refers to the gbrain repository's conventions and docs files for citation rules.
The skill depends on existing searchable pages and cached contradiction findings. Without a populated brain, results will be thin or empty.
- Tools required: search, query, get_page, list_pages, takes_search, find_contradictions, find_trajectory
- Repository files referenced: the gbrain conventions and docs for citation rules
- Content required: a populated brain with searchable pages
Trigger the skill with natural requests
You do not need a special command. Ask about one idea in plain language. The skill's triggers include phrases such as 'idea lineage', 'trace the lineage of this idea', 'how has my thinking about', 'what is my current version of' and 'where did this idea come from'.
Name one idea clearly. If the phrase could refer to several distinct ideas in your brain, the agent will ask you which one you mean before it writes the synthesis.
- Run idea lineage on founder-led sales.
- How has my thinking about compounding trust changed?
- What is my current version of this idea?
- Where did this idea come from, and what did I abandon along the way?
Read the answer and check the evidence
For a normal answer the agent uses a fixed set of headings: Current Live Version, Lineage, Reversals and Contradictions, Abandoned Branches, Related Concepts and Evidence Gaps. Each non-gap claim carries a source citation with a date and quote where the source has them.
Confidence is marked high, medium or low. Confidence is low when evidence comes from a single weak snippet, an undated page or a fuzzy semantic match. Undated evidence is labeled 'undated'. Where a bucket has no evidence, the answer says 'No clear evidence found' and notes what was checked.
- Current Live Version: 1-3 sentences with confidence
- Lineage: first mention, best articulation, turning point
- Reversals and Contradictions, Abandoned Branches, Related Concepts, Evidence Gaps
Safety notes: scripts and network
The skill does not run scripts and does not use the network, according to its listing. It works through the gbrain tools listed above and does not modify brain pages unless you explicitly ask for a saved artifact.
Because the skill is read-only by default, you can use it to inspect your notes without changing them. The skill also does not launch expensive contradiction probes on its own; it reads cached findings unless you ask for the heavier pass.
- Runs scripts: no
- Uses network: no
- Default mode: read-only
Tips for better lineage answers
Use the exact phrase you use in your notes, and say so if you have a variant. The skill searches for exact phrase variants and also runs one semantic query, so both help.
If the answer comes back thin, ask whether the idea is really a concept rather than a metric or status. Metric and status questions go to find_trajectory instead.
- Use the phrase as you wrote it in your notes.
- Name one idea per request.
- Ask for a saved page only after you have read the answer.
What people use it for
- Seeing when a belief first appeared in your notes and how your wording sharpened over time.
- Finding reversals in your stance on a topic, with before and after evidence.
- Checking what you abandoned along the way before you revisit an old idea.
- Tracing the origin of a concept across your own transcripts, meetings and articles for a writing project.
Frequently asked questions
How do I install the idea-lineage agent skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps you can instead upload it under Settings → Capabilities → Skills.
What do I need to use idea-lineage?
You need a gbrain knowledge base that exposes search, query, get_page, list_pages, takes_search, find_contradictions and find_trajectory, plus the gbrain repository's conventions and docs files. The skill also needs a populated brain, since results are thin without existing pages.
Is idea-lineage free and what is its licence?
The item is listed under the MIT licence. This page does not state pricing, so check the source listing for any cost details.
Does idea-lineage run scripts or use the internet?
According to its listing, it runs no scripts and uses no network. It works through the gbrain tools and is read-only by default.
How is idea-lineage different from concept-synthesis?
Idea-lineage traces one idea's evolution. Concept-synthesis builds a broad map of many concepts, so broad requests such as 'map all my concepts' are routed to concept-synthesis instead.
Why are my idea-lineage results thin or empty?
The skill depends on existing searchable pages and cached contradiction findings. If your brain is new or sparse, or the phrase does not match your notes, results will be limited and the answer will list evidence gaps.
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 idea-lineage if you keep a long-running gbrain knowledge base and want cited, evidence-backed answers about how your own thinking on one idea changed. Pick something else if you have no populated brain, want a broad concept map, or need to track numbers or job titles over time.
Download the skill