Sigil: Architecture Pattern Inventory

SkillMonitoring & ops

Use when: mapping a repository architecture and generating or refreshing a reusable architecture pattern inventory package with selective context, concept cards, dependency rules, testing alignment, and observability alignment.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Sigil: Architecture Pattern Inventory skill

What this skill tells your AI

The instructions your AI receives, as published by cyberalchemyai/arcanum in arcana/architecture-pattern-inventory/SKILL.md and read by ahel’s review.

  • creating a new architecture inventory package,
  • refreshing an existing architecture package from code evidence,
  • extracting reusable implementation patterns,
  • documenting layer and dependency rules,
  • creating concept and relationship cards,
  • aligning architecture guidance with testing and observability expectations,
  • preparing a codebase for agentic implementation with selective context loading.
  • repository root,
  • existing architecture docs,
  • source directories,
  • tests,
  • build or dependency configuration,
  • routing or module boundaries,
  • known architectural constraints,
  • desired output path,
  • whether existing package files should be updated or preserved.
architecture/

inside the target repository.

Recommended mapper roles:

  1. Structure mapper: identifies top-level folders, applications, packages, modules, and ownership boundaries.
  2. Layer mapper: identifies layers, dependency direction, allowed and forbidden imports, and boundary seams.
  3. Pattern inventorist: identifies repeated concepts, implementation motifs, naming conventions, and reusable cards.
  4. Testing mapper: identifies how architectural units are verified and where gaps exist.
  5. Observability mapper: identifies logs, metrics, traces, events, health checks, and operational signals tied to architecture.

Each mapper must return findings with evidence. The main agent synthesizes; mapper agents do not decide canonical package contents independently.

  • produce a navigable architecture package with an index and pattern-library entrypoint,
  • ground layer and pattern claims in repository evidence,
  • separate observed architecture from recommended architecture,
  • include dependency rules, testing alignment, and observability alignment,
  • make concept and relationship cards selectively loadable,
  • preserve existing package content unless update scope is explicit,
  • identify unresolved gaps rather than inventing missing architecture.
  • copying a generic architecture model over a repository without evidence,
  • treating folder names as architecture without checking usage and dependencies,
  • mixing product-specific terminology into a reusable package unless it is truly local to the target repo,
  • creating one giant architecture document when selective context files are needed,
  • letting mapper subagents write canonical files without synthesis,
  • documenting ideal rules as if they are currently enforced,
  • ignoring tests and observability when mapping implementation patterns.
## Architecture Pattern Inventory Result

- Repository: <path>
- Package path: <path>
- Mode: created | updated
- Mapper pass: subagents | local fallback
- Layers mapped: <summary>
- Patterns inventoried: <count and categories>
- Relationships inventoried: <count and categories>
- Files changed: <paths>
- Gaps deferred: <summary>
- Validation: <checks performed>
- Next refresh trigger: <condition>

Signals

GitHub stars
25
Forks
3
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
architecture-pattern-inventory
Source
github.com/cyberalchemyai/arcanum