JPeetz/agent-skills / skills/ai-agent-engineering/ai-agent-architecture
ai-agent-architecture
VerifiedUse this skill when designing, auditing, or diagnosing an AI agent's system architecture. Activates on "agent architecture", "agent design", "agent audit", "multi-agent system", "agent orchestration", "agent topology", "agent evaluation design", or "agent execution flow". Covers agent topology design, reference architecture patterns, audit frameworks, evaluation design, execution continuity, and architectural decision records. Outputs architecture diagrams, RFC-style design docs, audit reports with severity-ranked findings, and migration plans.
| File | Kind | Size |
|---|---|---|
| CHANGELOG.md | markdown | 1.7 KB |
| LICENSE | other | 1.0 KB |
| SKILL.md | skill md | 28.3 KB |
| evals/evals.json | config | 6.5 KB |
| references/sources.md | markdown | 7.2 KB |
CHANGELOG.md
View on GitHub# Changelog
All notable changes to this skill are documented here. Format follows
[Keep a Changelog](https://keepachangelog.com/en/1.1.0/); versioning follows
[Semantic Versioning](https://semver.org/spec/v2.0.0.html).
## [1.0.0] - 2026-11-14
### Added
- Initial release: `ai-agent-architecture` skill (new `ai-agent-engineering` category)
- Five topology patterns — single agent, supervisor, swarm, pipeline, mesh — with ASCII diagrams, use-when guidance, and the contract rules each implies
- Topology selection matrix keyed on design-time signals, with the rule "justify with a named pressure"
- Five-step architecture design process: scope → topology → contract → eval → iterate, each step producing an artifact
- Node-boundary contract table (input/output schema, error classes, idempotency key, timeout/retry, budget, version) and typed-envelope rule
- RFC-style design document skeleton
- Audit framework across four dimensions — completeness, security, observability, cost — with checklists, S1–S4 severity ranking, finding format, and report skeleton
- ADR template with context, decision, consequences, rejected alternatives, and signals that would reopen the decision
- Anti-pattern table with symptom, cause, and replacement for twelve common agent-architecture failures
- Evaluation design: five layers (unit, step, trajectory, end-to-end, online), golden-set construction rules, metrics table, pass-criteria discipline
- Execution continuity section: checkpointing, tested resume paths, partial results, compounding timeout budgets
- Red flags and dual verification checklists for completed designs and completed audits
- `evals/evals.json` with 4 trigger cases and 2 near-miss cases
- `references/sources.md` with upstream, vendor, and standards sources
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |