Agent Architecture

GitHubOfficialCreated by Stanislav BankovskiiDistributed by GitHub

Skill
Research and analysis
Agent tooling
MIT
No issues found

From the author

Design AI agent architectures through requirements discovery, or audit and diagnose architectural flaws in existing agents. Architecture only; excludes implementation and general code review.

— the description by Stanislav Bankovskii

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

This skill provides a systematic approach to defining the architecture of AI agents or auditing existing ones. It focuses on requirements discovery, structural integrity, and operational reliability, ensuring that agent designs are justified by specific needs rather than unnecessary complexity. It delivers comprehensive architectural packages, including diagrams and validation plans, without performing implementation or code-level reviews.

What it does

Designs new agent architectures based on clear requirements discovery.
Audits existing agent architectures to identify flaws and verify paths.
Diagnoses performance issues or failures in agent logic.
Produces architectural reports including Mermaid or C4 diagrams.
Defines validation loops, state management, and execution contracts.
Evaluates the necessity of complex components like RAG, memory, or multi-agent systems.

When to use it

When you need to design an AI agent from scratch and want a justified, efficient architecture.
When an existing agent is failing, producing errors, or not meeting performance goals.
When you need to verify if your current agent design is robust and scalable.
When you need to document an agent's architecture for stakeholders or developers.

Advantages

Promotes architectural simplicity by requiring justification for every added component.
Provides evidence-based audits rather than subjective assessments.
Ensures clear boundaries between agent logic, tools, and external effects.
Delivers professional, ready-to-use architectural documentation and diagrams.
Focuses on verifiable outcomes and robust error handling.

Usage examples

“I need to design an agent that handles customer support tickets. Can you help me define the architecture?”: The agent will guide you through a discovery process to define requirements, select the appropriate topology, and deliver an architecture package including diagrams and validation rules.
“My current agent keeps getting stuck in a loop when searching for data. Can you audit it?”: The agent will reconstruct the current execution paths, identify the root cause of the loop, and propose a corrected architectural approach with diagnostic checks.

Install

Download the skill as a ZIP:
https://autogineer.com/agent-skills/github/agent-architecture/download

Unzip it into the cross-client skills folder of your project:

mkdir -p .agents/skills && unzip agent-architecture.zip -d .agents/skills

Documentation by Stanislav Bankovskii

AI Agent Architecture

Help the user obtain a justified architecture for their task or an evidence-based audit of an existing agent. Deliver architectural decisions and ways to verify them, without implementing the agent. By default, completed work includes a PDF report and a visualization of the results. An “ideal architecture” fits the requirements, cost of failure, and team resources; it does not maximize the number of components.

Choose a route

Request	Route	Read
New agent, requirements are not yet clear	Design: working cases → early design → requirements and decision coverage → delivery	design.md, architecture-contract.md
Architecture from an existing specification	Design: fill in what is known and clarify only gaps	The same files; do not restart the interview
Review an agent already written	Audit: reconstruct actual paths → verify → deliver findings	audit.md, and architecture-contract.md as criteria
Agent makes mistakes, has degraded, or falsely reports “done”	Diagnosis within the audit: case → hypotheses → discriminating checks → correction and closure criterion	audit.md and diagnostic-review.md
Review and redesign	Audit first; its demonstrated problems become design inputs	audit.md first, then design.md

In either mode, read source-map.md once: it explains the origins of the principles and the textbook's limitations. The original PDF is not needed for ordinary skill use. scenarios.md is needed only to test the skill itself.

When choosing or revisiting the execution approach, use architecture-selection.md; when designing acceptance or reviewing quality claims, use evaluation-design.md. Develop the validation loop and completion evidence using validation-loop.md; for long-running/background work, pauses, recovery, and competing sessions, use execution-continuity.md, including storage, RTO/RPO, budgets, the human decision queue, and scheduling. Develop delegation, mutable memory, execution isolation, and long-running/streaming interaction only when the task has these properties. A section's existence does not make its question mandatory: material gaps under discovery-protocol.md determine depth.

Shared decision rules

First read the available specification, local instructions, architectural decisions, and relevant materials. Use code to reconstruct architecture, not to make unsolicited fixes. Do not run an application with external effects for an audit.
Maintain a brief register: source-confirmed / user requirement / proposal / assumption / open question / not applicable. Identify where requirements came from. A user decision and an architect's hypothesis have different statuses.
Corporate contracts and accepted decisions apply only within their own project. The textbook is an engineering reference, not a source of authority or a replacement for local canon. Identify conflicts rather than resolving them silently.
First consider ordinary automation without an LLM, a single call, and a predefined workflow. Introduce an agent loop, RAG, persistent memory, MCP, or multiple agents only for a concrete need. For each added complexity, identify its benefit, cost, verification method, and simpler alternative.
Do not select a model or framework before understanding the task. For a concrete selection, check current official documentation and version constraints. A documented capability is not yet demonstrated quality on the user's data.
Separate probabilistic model decisions from programmatically enforced rules. Describe where permissions, parameters, budget, and action admissibility are checked before an external effect, including bypass paths and resumption.
For a timed-out external write, a readback that finds nothing does not by itself prove that no effect occurred. Permit a retry only under an established downstream idempotency contract or authoritative proof of non-execution; otherwise retain effect unknown and reconcile or escalate. Apply this rule in concrete flows and examples as well as in the risk section.
An audit or design does not authorize writing code, changing agent settings, publishing, or initiating external actions. On a subsequent explicit implementation request, hand the architecture to the appropriate process; this skill does not continue into implementation itself.

How to work

Before an interview or audit planning, read discovery-protocol.md. Show a clear route and maintain a coverage map. By default, devote each turn to one decision or working episode; do not hide several independent topics inside one question. Material gaps and evidence determine depth. There is no fixed total round limit.

Deliver the first useful design as soon as context is sufficient, otherwise no later than the third answer; the count does not reset on continuation. This limits the wait for an early result, not the completeness of the interview. If the task is too unclear, show a map of what is understood and conditional options. After the sketch, continue investigating material gaps under the protocol; two or three rounds alone do not justify declaring readiness.

The first design includes the goal and boundaries, main capabilities and their outputs, recommended components, main flow and external actions, key constraints, assumptions, and open decisions. It is a sketch for early feedback. The interview budget limits the wait for a sketch, not design depth: develop it into an architecture package from what is already known, without waiting for a separate instruction to elaborate. If context suffices, deliver the package immediately. If the user explicitly asks only for a sketch, respect and label that depth.

Phrases such as “that's enough,” “let's go with this for now,” “the rest later,” or “enough questions” end requirements gathering: deliver the architecture from accumulated context in the same answer. Do not require a separate “now design it” instruction or end at “interview complete.” If a design has already been delivered, show its current final version or a substantive update. An explicit request to stop all work (“don't continue,” “that's all for today, stop”) means stop, rather than deliver a new design.

If the user does not know an answer, propose a justified option and label its status. Represent unknowns as assumptions and open decisions. Unclear authority blocks the corresponding external action in the proposed architecture, but not delivery of the architecture itself. Silence and ending the interview do not approve proposals.

After a significant answer, update the working summary of requirements and decisions. Save it in an agreed document if artifact creation is within the request; otherwise maintain it in the conversation. On continuation, start with that summary and changed information.

After the first design, clarify specific branches and uncovered material requirements, including real exceptions, human work, and feasibility. Explain which decision the answer will change; propose internal mechanisms yourself. Do not confine gap discovery to components already drawn or restart a questionnaire. Finish when the declared scope has sufficient coverage; if further confirmation is unavailable, deliver a conditional package with owners and checks for gaps.

Complete design with the architecture package from architecture-contract.md: domain capabilities and methods, output contracts, the structure of instructions/skills/materials, allocation between the existing platform and additions, a populated end-to-end example, and checks. Read capability-design.md for this part; in an audit, use it to check required capabilities. Describe the agent's main work deeply enough that a developer does not have to invent its method again. A platform name and a list of stages do not accomplish that.

Always cover limits on iterations, time, tokens/money, and tool calls, stopping rules, and what the user receives on stopping. Mark unknown values as open or proposed rather than inventing an agreed limit. An architecture package with skill specifications remains a design: it does not imply skill installation, code implementation, or verification of a running agent.

Complete an audit with demonstrated problems, separately identifying unknowns and accepted tradeoffs. Do not claim production readiness from reading code. Architectural readiness for implementation and demonstrated operational quality are different outcomes.

Final artifacts

When completing design, audit, or diagnosis, read result-delivery.md and create a PDF of the results with a rendered Mermaid or C4 diagram as appropriate; retain editable text and diagram source. Do this as part of completion without a separate user request to “make the PDF now.” An early sketch and intermediate answers do not require repeated export. Explicit user constraints (“chat only,” “no files/PDF”) and a request to stop all work take precedence. Creating the report does not authorize implementing or changing the reviewed agent.

Package metadata

This package is distributed under the MIT license. Optional client metadata supports compatible Agent Skills clients; Copilot uses SKILL.md and the linked references.

Files

ATTRIBUTION.mdAdded by AutoGineer982 B
LICENSE.txt1.1 KiB
SKILL.md9.8 KiB
agents/openai.yaml309 B
references/architecture-contract.md20.3 KiB
references/architecture-selection.md7.3 KiB
references/audit.md9.5 KiB
references/capability-design.md6.1 KiB
references/design.md10.2 KiB
references/diagnostic-review.md13.0 KiB
references/discovery-protocol.md8.6 KiB
references/evaluation-design.md7.5 KiB
references/execution-continuity.md13.4 KiB
references/result-delivery.md7.4 KiB
references/scenarios.md24.7 KiB
references/source-map.md18.4 KiB
references/validation-loop.md6.8 KiB

references/validation-loop.md

Validation Loop and Completion Evidence

Read when designing acceptance, release, and feedback, or diagnosing unconfirmed outcomes. These are functions of one architecture, not mandatory separate services. A simple task needs a compact description of applicable checks; explain non-applicability rather than creating empty chapters. The check set does not enlarge the interview or replace domain methods.

1. Function map

For each applicable function, populate input/version → property checked → mechanism and owner → execution timing → pass/fail/unknown → consequence → evidence. Show what the platform supplies, what is proposed, and what is unknown. “Not applicable” differs from a missing mandatory check.

Function	Architectural decision	Diagnostic check
Eval Registry	Versions of scenarios, oracle/rubric/scorer, run conditions, thresholds/admission criteria. Capabilities, regressions, escalations; long horizon when work is long-running	Which scenarios were actually executed/evaluated versus excluded; whether versions/denominators are comparable
Review	Who checks domain meaning, architecture, or artifact quality: program, model, human, or combination. Task-specific checklist and remediation bound	Whether the checker distinguishes good from plausible bad results and has enough primary data
Security Gate	Applicable asset-based threats/checks. For code: SAST/DAST, secrets/dependencies by risk; for text: data protection and output admissibility	Whether real paths are covered and dangerous effects/disclosure can precede checking
Policy Gate	Check concrete action, principal/tenant, resource, parameters, permissions, extra conditions; enforcement point	Parameter substitution, approval expiry/revocation, alternative handler, retry/resume; a negative verdict actually prevents action
Evidence Bundle Gate	Completeness, relevance, and freshness of evidence for the claimed outcome	All fields present but tests refer to an old version; effect unknown; links inaccessible or do not support claims
Release Gate	Admit a composite version to active use: owner/policy, checks, bounded pilot, stop/rollback	Whether an unverified version can activate; configuration/data/continuing-task state consistency
Telemetry Feedback	Operational signal → diagnostic case → evaluation/requirement → targeted change → comparison → accept/rollback	Whether failure becomes reproducible, results improve outside tuning, and criteria were not changed for a green score

Order depends on effects: write admission is checked before writing; result confirmation afterward. The diagram does not prescribe sequential execution of all seven components at every step. The registry may be a versioned test directory; review an existing check or a human; release for a local assistant a controlled version replacement. Platform documentation does not prove configuration or coverage.

Without a mandatory verdict, the result is not admitted. Choose a concrete outcome: waiting, stopping, or permitted degradation. Evaluator failure does not prove a domain-result error. Use evaluation-design.md and diagnostic-review.md for axes, measurement failures, and positive/negative controls.

2. Evidence Bundle

Design one verification package or manifest linking authoritative records. Choose format by environment; a separate service is unnecessary. For a material criterion, link:

task/run ID → R/E-ID → composite agent and material input versions → output ID/version → check/scorer and verdict → evidence link → completion status/unknown.

For actions, add operation identity, exact parameters/payload version, applicable authorization/policy decision, and confirmed effect state. For operations, add costs with accounting coverage and a link to subsequent user-value measurement. Composite version includes model, prompts/skills, tools, policies, runtime, and material data. Label open values explicitly.

Check:

evidence belongs to the required version, scenario, and data scope; PDF build date does not prove that;
links resolve for an authorized reviewer and content supports the conclusion; a hash proves identity, not correctness;
required checks actually executed; planned/skipped/error/unknown do not become pass;
actual effects remain separate from sent requests and intent; timeout preserves uncertainty until reconciliation;
package completeness, result correctness, action permission, and demonstrated benefit have separate statuses;
no secrets or unnecessary sensitive data are included; access, retention, and deletion follow applicable rules rather than universal permanent raw logging.

The model cannot grant itself final admission by writing done=true. For a proposed agent, the bundle specifies future evidence and a populated synthetic example, not real-run results. The final PDF under result-delivery.md presents findings and evidence links but does not replace evidence.

Example: experiment report

Synthetic run r18: requirement E4—a correct report for period P; input snapshot S3, method v2, output report R4. Numeric checking compares R4 with S3; domain review checks causal-inference limits. MAU changed from 1000 to 1100; no randomization, and prices also changed. The +10% calculation may pass while the experiment's contribution remains unconfirmed. The next-experiment decision rests on limitations, not an invented causal effect.

If R4 is replaced with R5, successful R4 checks do not admit R5. If R5 publication times out, the final status is “report prepared; publication unconfirmed,” with operation ID and the next authorized reconciliation step. Separately record the reason if publication was never authorized.

3. The checking component also has a contract

If a Guardian/Review-Agent is planned, distinguish observation / pre-effect interception / correction. For each mode, define visible data, intervention point, authority, timeout, unavailable behavior, cost, and its own false-accept/false-block checks. Correction is a separate action subject to normal admission rules; reviewer authority does not automatically expand.

Example: a write completes at t15; Guardian returns deny at t20. The monitor detected the event, but pre-write prohibition was not enforced. To verify a correction, send prohibited and permitted operations through the actual dispatcher/adapter: the former never reaches an external effect; the latter reaches it and is confirmed. Test Guardian unavailability separately. A model judge does not replace mandatory authorization or guarantee attack absence.

Review is bounded by agreed rounds, time/cost, and stopping rules. Unresolved disagreement becomes a specific open decision or escalation, not endless exchanges between worker and reviewer. Do not disable mandatory controls for speed or a diagnostic experiment.

Requirements and compatibility

Written for: GitHub Copilot

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Security

No issues found by our static scan· No issues found

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by

Stanislav Bankovskii

Publisher

GitHub (Official)

Distributed by

GitHub

Source

github/awesome-copilot · skills/agent-architecture · commit 797337f

Licence

MIT License (redistribution allowed)

ATTRIBUTION.md (written by AutoGineer)

Synced

2026-11-09

AutoGineer is not affiliated with or endorsed by Stanislav Bankovskii or GitHub.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download agent-architecture.zip