understand
Skill1 starsChanged 3 months agoMITView on GitHub
Use when the user wants to learn or research a topic. Map relevant dimensions, set depth by the learning goal, and explore the requested direction; works independently of BestWay.
18 other projects use a version of this.
Adopted by
Most recently taken up by
- thoughtbot/rails-consultanttheir own version, first seen Nov 8, 2026
- taljeon/plzdotheir own version, first seen Nov 1, 2026
- stevepolitodesign/skillstheir own version, first seen Oct 14, 2026
- jayden-dang/skillstheir own version, first seen Oct 13, 2026
- fcescob/understandtheir own version, first seen Oct 13, 2026
- bjcoombs/ai-native-toolkittheir own version, first seen Oct 13, 2026
Used with
Projects with it also use understand-chat, understand-dashboard and 5 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
karanmrnSkillunderstand-chat
Use when you need to ask questions about a codebase or understand code using a knowledge graph
Also in 3 projects that use this
karanmrnSkillunderstand-dashboard
Launch the interactive web dashboard to visualize a codebase's knowledge graph
Also in 3 projects that use this
atopileSkillgraph
How the Zig-backed instance graph works (GraphView/NodeReference/EdgeReference), the real Python API surface, and the invariants around allocation, attributes, and cleanup. Use when working with low-level graph APIs, memory management, or building systems that traverse the instance graph.
Also in 3 projects that use this
AWSSkillship
Finalize dev work on an issue branch — update docs, push, and open a PR.
Also in 3 projects that use this
Google CloudSkilltriage
Gather context on a change or bug and plan the implementation.
Also in 3 projects that use this
- +2 more from karanmrn
Other versions18 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 18
- smallnest/goal-workflow★ 283same name
- bjcoombs/ai-native-toolkit★ 32same name
- thoughtbot/rails-consultant★ 29same name
- agoodway/.claude★ 11same name
- stevepolitodesign/skills★ 10same name
- taljeon/plzdo★ 9same name
- fcescob/understand★ 4same name
- digitaldreams/tuhin★ 3same name
- jkjackson2000/understand★ 2same name
- jayden-dang/skills★ 1same name
8 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository28 other files it gives its agents.
- add-commentSkill
- agent-supervisorSkill
- ai-seoSkill
- anti-goalsSkill
- bestwaySkill
- codebase-renderSkill
- competitor-profilingSkill
- copywritingSkill
- decision-reviewSkill
- directory-submissionsSkill
- explain-and-saveSkill
- happy-featSkill
- i-have-adhdSkill
- investment-researchSkill
- launchSkill
- learning-new-languageSkill
- lock-decisionsSkill
- moleSkill
- onboardingSkill
- open-source-reviewSkill
- pmSkill
- product-marketingSkill
- product-startupSkill
- prototype-treeSkill
- read-sessionSkill
- reverse-engineer-marketingSkill
- socialSkill
- update-harnessSkill
The file
We split this skill into the rules it gives an agent
7 rules. 0 are common, 7 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 16 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 13
Get grounded knowledge: Read relevant project or domain material.
- Distinctive: line 23
Use a compact map only when the topic has several meaningful branches.
- Distinctive: line 27
Do not quiz the user or force recall exercises.
- Distinctive: line 28
Keep privacy and other existing boundaries intact.
- Distinctive: line 34
When the user wants to understand a topic, map its dimensions and deepen the chosen branch here.
- Distinctive: line 35
When the user shifts from understanding to choosing or planning an action, BestWay may be used on its own or with this skill's findings.
- Distinctive: line 36
If BestWay needs an unfamiliar fact, research only the decision-relevant part.
“Common” is measured: kept by at least 6 in 10 of the 16 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Understand a topic
Use this skill for learning and research requests. It can complete a useful explanation on its own; BestWay is an optional follow-up when the user wants to choose or plan an action. Do not apply it to unrelated execution requests or turn every question into a course.
Workflow
- Set the learning goal. Infer what the user wants to understand from the request when clear. Ask one short question only when materially different directions are plausible and the answer would change the explanation. Do not ask for motivation unless it affects the scope.
- Get grounded knowledge. Read relevant project or domain material. For unfamiliar, changing, or consequential claims, check authoritative sources or direct evidence. Separate verified facts, inference, and unknowns. Skip tangents.
- Map breadth before depth. Identify the dimensions that matter to this subject and goal. Depending on the topic, these may include purpose, participants and interests, concepts, mechanisms, lifecycle, boundaries, trade-offs, failure modes, or practical use. Do not force every subject into the same checklist.
- Set a goal-relative depth for each dimension. When a map is useful, show each dimension's main question, what it explains, depth reached, what would be enough for the current goal, and the next frontier. Include participants or stakeholder roles when they help explain the subject. Depth may move from orientation to structure, mechanisms, edge cases, or underlying theory; these are optional layers, not mandatory levels.
- Follow the chosen branch. Explore the direction the user chose. Keep the broader map visible so they can see adjacent dimensions and choose where to go next.
- Check the model. Trace a concrete example through the concepts and relationships. Compare the explanation with evidence; revise it if they disagree. Mark unresolved points instead of guessing.
- Teach in layers. Start with a plain definition, explain essential terms, then add a small example or diagram when it makes the relationships clearer. Label invented examples as pseudo-examples. For several interacting parts, build the diagram in readable stages rather than overwhelming the user with every detail at once.
- Write knowledge only on request. Do not create or update files just because research was performed. When the user explicitly asks to save the knowledge, find the canonical project or domain document, update it instead of creating a duplicate, and keep sources attached to researched claims.
Mapping and stopping
- Use a compact map only when the topic has several meaningful branches. Useful columns include dimension, what it explains, depth reached, enough for this goal, and next frontier.
- A dimension rarely has one absolute endpoint. Stop the chosen branch when it is deep enough for the stated goal and important claims are supported. Mark deeper layers that remain unexplored; do not imply they were covered.
- A useful answer may explore one branch deeply while leaving others at overview level. Do not expand every branch equally unless the user asks for that.
- For a narrow factual learning question, answer directly and map only the context needed to make the answer understandable.
- Do not quiz the user or force recall exercises. Do not turn a question into a course unless requested.
- Keep privacy and other existing boundaries intact. A learning request does not by itself authorize saving personal information or changing an external system.
Relationship with BestWay
This skill builds a knowledge model and can be used without BestWay. BestWay is an optional, independent skill for choosing an execution route.
- When the user wants to understand a topic, map its dimensions and deepen the chosen branch here. Do not force a route comparison.
- When the user shifts from understanding to choosing or planning an action, BestWay may be used on its own or with this skill's findings.
- If BestWay needs an unfamiliar fact, research only the decision-relevant part. Pass along concise verified facts, assumptions, and unknowns; do not repeat the full explanation or require one skill to precede the other.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

