architecture-doc
Skill0 starsChanged 3 months agoMITView on GitHub
When asked to set up or write a project's architecture document — produce a project-root ARCHITECTURE.md by first grilling the user on the architecture decisions. Use for "write the architecture doc", "create an ARCHITECTURE.md", "settle the project structure", or planning a new codebase's layout. Skip for fixing bugs, reviewing existing docs, or writing code.
2 other projects use a version of this.
Adopted by
Most recently taken up by
- sterlp/eclipse-peon-aitheir own version, first seen Nov 1, 2026
- arsudsandesh97/Revoacttheir own version, first seen Jul 22, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- sterlp/eclipse-peon-ai★ 32same name
- arsudsandesh97/Revoact★ 3same name
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 repository6 other files it gives its agents.
- architecture-driven-codingSkill
- code-reviewerSkill
- create-skillSkill
- grillingSkill
- product-builderSkill
- test-case-writerSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 23
When the user confirms shared understanding, write ARCHITECTURE.md to the project root using the format in assets/architecture-template.md.
- Very specific: line 46
Follow the section order and headings in assets/architecture-template.md exactly — the coding skill reads this file to map features to …
- Very specific: line 51
If ARCHITECTURE.md already exists, read it first, then grill, then rewrite
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Instructions
- Load the grilling skill with the Skill tool and apply its discipline: ask one question at a time, give a recommended answer for each decision, look up facts from the environment instead of asking, and do not write anything until the user confirms shared understanding.
- Scan the project first — working directory, project root, any existing
CONVENTIONS.md,README.md,package.json, or config files. Extract facts yourself; grill only on decisions. - Walk the decision tree below, in order. Ask one question at a time and wait for each answer before continuing. For every question, provide your recommended answer.
- When the user confirms shared understanding, write
ARCHITECTURE.mdto the project root using the format inassets/architecture-template.md.
Decision tree (grill topics)
Grill the user on each topic. Sections 6 and 7 are optional — only walk them if applicable; otherwise confirm they don't apply and move on.
- Overview — what the system does, its goals and non-goals
- Tech stack — language, framework, key libraries, tooling, versions
- Directory structure — layout style (feature-based, layered) and rationale
- Module / component breakdown — how responsibility splits across modules
- Data flow — how data moves between modules and layers
- Contracts & interfaces (optional) — only when cross-module or external contracts are worth pinning down
- Boundaries & rules (optional) — only when the architecture enforces hard rules, e.g. "UI never queries the database"
- Conventions reference — which
CONVENTIONS.mdapplies and any deviations (read from the filesystem, don't grill unless ambiguous) - Open questions — record anything left unresolveable at write time
Output rules
- Follow the section order and headings in
assets/architecture-template.mdexactly — the coding skill reads this file to map features to modules, so heading names must stay stable - Sections 1–5, 8, and 9 are mandatory
- Sections 6 and 7 are optional — omit them entirely if not applicable; no empty stubs
- If
ARCHITECTURE.mdalready exists, read it first, then grill, then rewrite - Report afterwards: where the file lives, and which sections were omitted as not applicable
Example
Input: "Write the architecture doc for our new analytics platform."
- Load the grilling skill via the Skill tool.
- Scan the repo — read
README.md,CONVENTIONS.md,package.json. - Grill one question at a time, each with a recommendation: "What does the platform do at its core?" (→ track events, run dashboards) — "Recommended stack?" (→ Next.js + ClickHouse) — wait for each answer.
- Once understanding is confirmed, write
ARCHITECTURE.mdat the project root — all 9 sections filled, except 6/7 which are omitted if the user says no external contracts or hard boundaries exist.
Out of scope
- Editing an existing architecture doc in place — edit
ARCHITECTURE.mddirectly - Writing code per an architecture — the architecture-driven-coding skill enforces it
- Reviewing a doc you did not produce
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.

