Agent skill · product strategy · aiskillstore
architectural-forensics
Master protocol for deconstructing agent frameworks to inform derivative system architecture. Use when (1) analyzing an agent framework's codebase comprehensively, (2) comparing multiple frameworks to select best practices, (3) designing a new agent system based on prior art, (4) documenting architectural decisions with evidence, or (5) conducting technical due diligence on AI agent implementations. This skill orchestrates sub-skills for data substrate, execution engine, cognitive architecture, and synthesis phases.
Why this skill is useful
Adds a comprehensive protocol with executable scripts for analyzing agent frameworks and generating architectural outputs that extend the AI's capabilities.
What it needs
Requires python installed locally. About 4k tokens when loaded. Last updated 2026-08-07. 405 stars on the source repository.
What this skill does
Architectural Forensics Protocol Deconstruct agent frameworks to inform derivative system architecture. Mission Distinguish between software engineering decisions (how it runs) and cognitive architecture decisions (how it thinks) to extract reusable patterns for new systems. Quick Start Protocol Phases Phase 1: Engineering Chassis Analyze the software substrate. See references/phase1-engineering.md for detailed guidance. Analysis Focus Files Output ---------- ------------- -------- Data Substrate types.py, schema.py, state.py Typing strategy, mutation patterns Execution Engine runner.py, executor.py, agent.py Async model, control flow topology Component Model base.py, interfaces.py Abstraction depth, DI patterns Resilience executor.py, try/except blocks Error propagation, sandboxing Phase 2: Cognitive Architecture Extract agent "business logic". See references/phase2-cognitive.md for detailed guidance. Analysis Focus Files Output ---------- ------------- -------- Control Loop agent.py, loop.py Reasoning pattern, step function Memory memory.py, context.py Context assembly, eviction policies Tool Interface tool.py, functions.py Schema generation, error feedback Harness-Model Protocol llm.py, adapters/, stream.py Wire format, tool call encoding, agentic primitives Multi-Agent orchestrator.py, router.py Coordination model, state sharing Phase 3: Synthesis Generate actionable outputs: 1. Best-of-Breed Matrix → Framework comparison table 2. Anti-Pattern Catalog → "Do Not Repeat" list 3. Reference Architecture → New framework specification Execution Workflow Quick Analysis (Single Framework) For rapid assessment, run the minimal path: Output Directory Structure State Management & Resumption The protocol is designed to be stateful and resumable. Idempotency: The Orchestrator tracks progress in manifest.json and will skip frameworks marked as completed. Clean Slate Resumption: If a run is interrupted, frameworks marked as inprogress are considered "stale". …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills aiskillstore/architectural-forensics