13-Layer Architecture
Understanding the core execution hierarchy of Synara.
Agent Domain Workflows
Select a domain below to explore the interactive agent interactions and pipelines.
Loading interactive diagram...
Synara is deeply inspired by modern robust software architectures, separating responsibilities into distinct layers. This ensures that when an AI Agent fails, it fails predictably and gracefully.
Layer Details
| Layer | Component | Description |
|---|---|---|
| 1 | .synara/agents/ | The core intelligence. System prompts defining roles, boundaries, and step-by-step thinking processes. |
| 2 | .synara/skills/ | Specialized capabilities (e.g., CLI parsing, Android integration) dynamically injected into the agent's context. |
| 3 | .synara/playbooks/ | Standardized workflows guiding agents through multi-step, complex procedures. |
| 4 | .synara/quality_gates/ | Automated strict verification gates before code is committed. |
| 5 | .synara/workflows/ | Structured multi-agent pipelines. |
| 6 | .synara/execution-engine/ | Runners, CLI wrappers, and execution reports. |
| 7 | .synara/quality_gates/ | Output validator schemas and lint rules. |
| 8 | .synara/memory-store/ | ADR histories, lessons learned, and incidents. |
| 9 | .synara/feedback-loop/ | Runtime metrics, anomalies, and alert configurations. |
| 10 | .synara/knowledge/ | Domain wikis and technical documentation. |
| 11 | .synara/governance/ | Audit trails, compliance checklists, and scope constraints. |
| 12 | .synara/metrics/ | Token pricing and usage limits. |
| 13 | .synara/templates/ & .synara/rules/ | Houses templates and active ruleset policies. |
Checksum-Based File Protection
Synara utilizes a robust file-writing constraint (safe_write_file) powered by SHA-256 Checksums. This prevents an AI agent from accidentally wiping out manual code you just wrote.