원클릭으로
add-agent
Scaffold a new Agno agent with boilerplate, register in main.py, and add config.yaml entry
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Scaffold a new Agno agent with boilerplate, register in main.py, and add config.yaml entry
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Scaffold eval test cases for an agent in backend/evals/test_cases.py, following the TestCase dataclass pattern used by DATA_AGENT_CASES, KNOWLEDGE_AGENT_CASES, and WEB_SEARCH_CASES
Verify all project documentation reflects current codebase state after feature changes
Scaffold a new Agno multi-agent team with boilerplate, register in main.py, and create docs page
Scaffold a new Agno workflow with Steps, Loops, and Conditions, register in main.py, and create docs page
Scaffold a new mise task with correct headers, flags, and conventions
Pre-release validation checklist — verify versions, CI, docs, and Docker before running mise run release
| name | add-agent |
| description | Scaffold a new Agno agent with boilerplate, register in main.py, and add config.yaml entry |
| disable-model-invocation | true |
Create a new Agno agent and integrate it into Apollos AI.
Ask the user for:
Create backend/agents/{id_with_underscores}.py following this pattern:
from agno.agent import Agent
from agno.guardrails import PIIDetectionGuardrail, PromptInjectionGuardrail
from backend.db import get_postgres_db
from backend.models import get_model
{variable_name} = Agent(
id="{agent-id}",
name="{Agent Name}",
model=get_model(),
db=get_postgres_db(),
instructions=["{description}"],
pre_hooks=[PIIDetectionGuardrail(mask_pii=False), PromptInjectionGuardrail()],
enable_agentic_memory=True,
enable_session_summaries=True,
add_datetime_to_context=True,
num_history_runs=5,
markdown=True,
)
Key conventions:
get_model() from backend/models.py (never inline LiteLLMOpenAI)db=get_postgres_db() for persistent statepre_hooks) — this is mandatory for all agentsAdd the import and include in the agents list:
from backend.agents.{module_name} import {variable_name}
agent_os = AgentOS(
agents=[knowledge_agent, mcp_agent, {variable_name}],
...
)
Add a quick-prompts entry in backend/config.yaml for the new agent.
If the agent uses tools that require new packages, add them:
uv add <package>
Create docs/agents/{id}.mdx with frontmatter, code example, features, and example queries.
Add the page to docs/docs.json navigation under the Agents group.
Update these files to reflect the new agent:
README.md — Agents tablePROJECT_INDEX.md — Agents section and core modules.serena/memories/project-overview.md — Agents listdocs/agents/overview.mdx — Agents table and card groupTell the user:
backend/agents/{name}.pybackend/main.pybackend/config.yamldocs/agents/{id}.mdxmise run docker:up or docker compose restart