architecture-blueprint
Skill1 starsChanged 5 months agoMITView on GitHub
Maintains an Architecture Blueprint so builders track what agents built across sessions. In Continuum repos, defer to X-3 (`.continuum/blueprint/`, BlueprintPanel) — do not duplicate. In other repos, use markdown `ARCHITECTURE.md` + append-only change feed. Invoke explicitly when the user mentions schema drift, lost track of architecture, what agents built, mental model sync, or architectureImpact.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- codexskills/agent-forgetheir own version, first seen Oct 17, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- codexskills/agent-forge★ 2same 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 repository33 other files it gives its agents.
- burnSkill
- calm-ui-auditSkill
- capitalizeSkill
- cardSkill
- commit-mineSkill
- continuity-contextSkill
- elonizeSkill
- frontierSkill
- full-throttleSkill
- grade-sessionSkill
- groundSkill
- harmonizeSkill
- harvestSkill
- honeSkill
- milkSkill
- mirrorSkill
- moonwalkSkill
- nightshiftSkill
- page-auditSkill
- pmfSkill
- postSkill
- promptSkill
- publishSkill
- razeSkill
- redlineSkill
- relaySkill
- snipSkill
- stashSkill
- tasteSkill
- tidySkill
- untilSkill
- vacationSkill
- vaultSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 7 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
Continuum guard: If this is a Continuum repo, use X-3 and blueprint_append_agent / continuum-report — do not create parallel …
- Distinctive: line 31
Before/after large refactors or agent swarms
- Distinctive: line 60
When closing a substantial task, emit one report block in the assistant message (user can copy to feed).
- Distinctive: line 86
Only when user asks. Pick 3 recent change rows, hide layer names, user guesses layer, reveal mapping.
- Distinctive: line 91
No numeric drift scores, leaderboards, or red "failed" gates in the feed
- Distinctive: line 92
No regex-parsing framework internals for discovery (use directory conventions)
- Distinctive: line 93
Do not edit the plan file if user attached an architecture plan
- Distinctive: line 94
Keep stable IDs/path globs when refreshing so change feed links stay valid
≈ “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
Architecture Blueprint
Goal
Answer: What is this system, and what changed since I last looked? — without a graph DB, fifth surface, or topology quiz gate.
Repo detection (read first)
| Repo | Primary artifact | Agent report block |
Continuum (has .continuum/ or BlueprintPanel) | X-3: .continuum/blueprint/ + BlueprintPanel UI | continuum-report JSON with "architectureImpact": ["path/..."] |
| Any other repo | ARCHITECTURE.md + ARCHITECTURE-CHANGES.md | architecture-report fenced block |
Continuum guard: If this is a Continuum repo, use X-3 and blueprint_append_agent / continuum-report — do not create parallel ARCHITECTURE.md unless the user explicitly asks for markdown-only dogfood.
When to use
- Multi-day agent-heavy work on one repo
- User can't explain module boundaries from memory
- Before/after large refactors or agent swarms
- User explicitly invokes this skill or asks for architecture tracking
Workflow (generic repos only)
Skip steps 2–3 in Continuum — append via X-3 instead.
- Scan once — list top-level layers (≤6): shell, surfaces/modules, backend/crates, IPC/API, docs, automation. Cap 7 visible items per layer; summarize overflow as "+N more".
- Write
ARCHITECTURE.md— one-line project summary + layer sections with plain-language bullets (no jargon labels like "god mode"). - Create
ARCHITECTURE-CHANGES.md— append-only feed:
## 2026-07-06 — git — feat: add auth module
- Summary: Committed changes in Backend, Surfaces.
- Paths: src/auth.ts, crates/session/...- On task completion — append agent-sourced row (no commit required):
## 2026-07-06 — agent — IPC wiring for bookmarks
- Summary: Added bookmark IPC handlers and UI hooks.
- Paths: src-tauri/src/commands.rs, ...- Optional git hook — only if user asks; append commit rows with changed paths + first line of message.
Agent report contract
When closing a substantial task, emit one report block in the assistant message (user can copy to feed).
Continuum repos — prefer this block:
```continuum-report
architectureImpact:
- src/surfaces/Build/BuildSurface.tsx
summary: Added file watcher debounce to indexer pipeline.
```Generic repos — use this block:
```architecture-report
impact_paths:
- src/surfaces/Build/BuildSurface.tsx
summary: Added file watcher debounce to indexer pipeline.
```Append a matching row to ARCHITECTURE-CHANGES.md (or X-3 feed in Continuum) only when impact paths are non-empty.
Mental model sync (optional, ≤3 min)
Only when user asks. Pick 3 recent change rows, hide layer names, user guesses layer, reveal mapping. Never block further work on pass/fail.
Constraints
- Markdown first in generic repos — no default web UI, xyflow, or LSIF graph
- No numeric drift scores, leaderboards, or red "failed" gates in the feed
- No regex-parsing framework internals for discovery (use directory conventions)
- Do not edit the plan file if user attached an architecture plan
- Keep stable IDs/path globs when refreshing so change feed links stay valid
Adversarial self-check (before finishing)
- Did I detect repo type and pick X-3 vs markdown correctly?
- Could a tired builder get oriented in <60s from text feed alone?
- Every change row names ≥1 touched path?
- Did I avoid building Continuum IPC when markdown or X-3 already suffices?
- Layer count ≤7 per section?
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.

