Ejecuta cualquier Skill en Manus
con un clic

Ejecuta cualquier Skill en Manus con un clic

test-agent-workflow

Estrellas3

Forks0

Actualizado2 de marzo de 2026 a las 00:37

Test an agentic system end-to-end. Use when verifying agent routing, subagent delegation, handoff payloads, compliance gates, and memory persistence after building agents and MCP servers.

Instalación

Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.

Ejecutar en Manus

Fuente

appliedcognetics

appliedcognetics/Swimwear

Abrir repositorio de GitHub Ver repositorios del creador

Descarga

Ejecutar en Manus

Ocupaciones relacionadasSOC

Basado en la clasificación ocupacional SOC

Analistas de garantía de calidad de software y probadoresOcupaciones informáticas y matemáticas·SOC 15-1253

SKILL.md

readonly

name	test-agent-workflow
description	Test an agentic system end-to-end. Use when verifying agent routing, subagent delegation, handoff payloads, compliance gates, and memory persistence after building agents and MCP servers.
argument-hint	Agent or workflow to test (e.g., "full pipeline", "compliance gate", "research agent")

Test Agent Workflow

When to Use

After creating or modifying agents — verify they respond correctly
After wiring MCP servers — verify agents can call tools
After changing handoff logic — verify subagent delegation still works
Before demo or release — full end-to-end validation
Debugging a broken workflow — isolate which agent or handoff fails

Procedure

Ask the user what to test: specific agent, specific handoff, or full end-to-end pipeline
Read the test scenarios for pre-built test cases
For the selected scope, execute each test scenario:
- Agent isolation test: Invoke agent directly with sample input, verify output format matches spec
- Tool connectivity test: Verify MCP server tools return valid responses
- Handoff test: Invoke a caller agent, verify it delegates to the correct subagent with correct payload
- Compliance gate test: Submit non-compliant data, verify it gets flagged and blocked
- Memory persistence test: Run a workflow, verify entries appear in expected /memories/repo/ paths
For each test, record: PASS/FAIL, input used, output received, expected output
Report results using the output format below
For failures: identify the broken component (agent body, tool handler, handoff payload, MCP config) and suggest a fix

Validation Checklist

All agents respond to direct invocation
All MCP server tools return valid JSON
Handoff payloads match the shapes defined in authoring-rules.md
Compliance agent correctly flags non-compliant inputs
Memory entries have proper attribution metadata
End-to-end pipeline completes without errors
Output format matches agent specs

Output

Test	Scope	Input	Expected	Actual	Status
{test name}	{agent/handoff/e2e}	{sample input}	{expected output shape}	{actual output}	PASS/FAIL

Más de este repositorio

mismo repositorio

bootstrap-agentic-system

appliedcognetics/Swimwear

One-shot workspace setup for a new agentic system. Use when starting a new project that needs agents, skills, MCP servers, and structured knowledge. Creates full directory structure, copilot-instructions.md, tool registry, and memory conventions.

2026-03-023

create-agent

appliedcognetics/Swimwear

Create a new .agent.md file for the investment partner agentic system. Use when building a new agent, adding a subagent, or scaffolding agent definitions. Guides through template, authoring rules, tool selection, and handoff design.

2026-03-023

create-mcp-server

appliedcognetics/Swimwear

Scaffold a local MCP server with synthetic data for agentic system development. Use when creating a new data source, building tool endpoints, or generating test data for agents to consume.

2026-03-023

create-skill

appliedcognetics/Swimwear

Scaffold a new SKILL.md file with references and assets directories. Use when building a reusable skill for any agentic system. Guides through naming, procedure design, reference file creation, and validation.

2026-03-023

memory-conventions

appliedcognetics/Swimwear

Guide for storing and retrieving institutional knowledge using Copilot built-in memory. Use when an agent needs to persist decisions, research, audit trails, or preferences to the correct memory scope with proper attribution.

2026-03-023

Test Agent Workflow

When to Use

After creating or modifying agents — verify they respond correctly

After wiring MCP servers — verify agents can call tools

After changing handoff logic — verify subagent delegation still works

Before demo or release — full end-to-end validation

Debugging a broken workflow — isolate which agent or handoff fails

Procedure

Ask the user what to test: specific agent, specific handoff, or full end-to-end pipeline

Read the test scenarios for pre-built test cases

For the selected scope, execute each test scenario:

Agent isolation test: Invoke agent directly with sample input, verify output format matches spec
Tool connectivity test: Verify MCP server tools return valid responses
Handoff test: Invoke a caller agent, verify it delegates to the correct subagent with correct payload
Compliance gate test: Submit non-compliant data, verify it gets flagged and blocked
Memory persistence test: Run a workflow, verify entries appear in expected /memories/repo/ paths

For each test, record: PASS/FAIL, input used, output received, expected output

Report results using the output format below

For failures: identify the broken component (agent body, tool handler, handoff payload, MCP config) and suggest a fix

Validation Checklist

All agents respond to direct invocation

All MCP server tools return valid JSON

Handoff payloads match the shapes defined in authoring-rules.md

Compliance agent correctly flags non-compliant inputs

Memory entries have proper attribution metadata

End-to-end pipeline completes without errors

Output format matches agent specs

Output

Test

Scope

Input

Expected

Actual

Status

{test name}

{agent/handoff/e2e}

{sample input}

{expected output shape}

{actual output}

PASS/FAIL