Agently Docs

EN

中文

EN

中文

Appearance

Official Agently Skills for Coding Agents ​

Agently-Skills is the official installable skills package for Codex / Claude Code and similar coding agents. It replaces the old agent_docs.zip flow, copied doc packs, and earlier local demo skills.

Official repository:

Official docs:

1. What the current Skills model is ​

The current Agently-Skills model is not just an API snippet pack. It teaches a coding agent to work along Agently-native capability boundaries:

  • recognize whether a request belongs to the request side, extension side, or orchestration side
  • choose the right skill or skill combination before implementation
  • prefer native Agently surfaces instead of starting with custom wrappers, parsers, or workflow glue
  • organize settings, prompts, tools, workflows, and runtime artifacts into a maintainable project shape

The goal is not shallow snippet generation. The goal is a complete project that actually fits Agently.

2. Correct setup path ​

Recommended flow:

  1. Install the official Agently Skills through the skills mechanism supported by your coding agent.
  2. Keep AGENTS.md / CLAUDE.md focused on repo-specific rules only.
  3. Let the installed official skills provide Agently knowledge and routing instead of vendoring a full Agently docs pack into each repo.

AGENTS.md / CLAUDE.md should now mostly contain:

  • directory layout and module boundaries
  • build, test, and release requirements
  • local coding conventions
  • internal docs that must be read first

They should no longer say:

  • "download and unzip agent_docs.zip"
  • "copy the whole Agently docs pack into this repo"
  • "prefer the old local mock skills"

3. Routing model ​

Use this current routing model when deciding where to start:

  • business goal, product behavior, project refactor, or unclear owner layer: agently-playbook
  • provider wiring, env vars, or model settings separation: agently-model-setup
  • prompt structure, prompt config, YAML/JSON prompt behavior, or mappings: agently-prompt-management
  • stable structured fields, required keys, or machine-readable output: agently-output-control
  • reuse one model result, access text/data/meta, or consume streaming output: agently-model-response
  • session continuity, restore, or memory: agently-session-memory
  • tools, MCP, FastAPIHelper, auto_func, or KeyWaiter: agently-agent-extensions
  • embeddings, vector indexing, retrieval, or KB-backed answers: agently-knowledge-base
  • TriggerFlow orchestration, runtime stream, event-driven fan-out, or mixed sync/async workflow: agently-triggerflow
  • LangChain / LangGraph migration: start with agently-migration-playbook, then route to the matching migration leaf

The key rule is: prefer Agently-native surfaces before inventing custom wrapper layers.

4. Current public catalog ​

According to the current official repository, the public catalog contains 12 skills:

Entry ​

  • agently-playbook Top-level router for unresolved product, assistant, automation, workflow, evaluator, or project-structure refactor requests.

Request Side ​

  • agently-model-setup Provider connection, dotenv-based settings, transport setup, and settings-file-based model separation.
  • agently-prompt-management Prompt composition, prompt config, YAML-backed prompt behavior, mappings, and reusable request-side prompt structure.
  • agently-output-control Output schema, field ordering, required keys, and structured output reliability.
  • agently-model-responseget_response(), parsed results, metadata, streaming consumption, and response reuse.
  • agently-session-memory Session-backed continuity, memo, restore, and request-side conversational state.

Request Extensions ​

  • agently-agent-extensions Tools, MCP, FastAPIHelper, auto_func, and KeyWaiter.
  • agently-knowledge-base Embeddings plus Chroma-backed indexing, retrieval, and retrieval-to-answer flows.

Workflow ​

  • agently-triggerflow TriggerFlow orchestration, runtime state, runtime stream, workflow-side model execution, event-driven fan-out, and mixed sync/async orchestration.

Migration ​

  • agently-migration-playbook Top-level migration router for LangChain or LangGraph systems.
  • agently-langchain-to-agently Direct LangChain agent-side migration guidance.
  • agently-langgraph-to-triggerflow Direct LangGraph orchestration migration guidance.

5. Common install paths ​

Install the full official skills repository:

bash
npx skills add AgentEra/Agently-Skills

If you want the smallest starting point:

bash
npx skills add AgentEra/Agently-Skills --skill agently-playbook

Common bundles:

request-core

bash
npx skills add AgentEra/Agently-Skills --skill agently-playbook
npx skills add AgentEra/Agently-Skills --skill agently-model-setup
npx skills add AgentEra/Agently-Skills --skill agently-prompt-management
npx skills add AgentEra/Agently-Skills --skill agently-output-control
npx skills add AgentEra/Agently-Skills --skill agently-model-response

request-extensions

bash
npx skills add AgentEra/Agently-Skills --skill agently-playbook
npx skills add AgentEra/Agently-Skills --skill agently-agent-extensions
npx skills add AgentEra/Agently-Skills --skill agently-session-memory
npx skills add AgentEra/Agently-Skills --skill agently-knowledge-base

workflow-core

bash
npx skills add AgentEra/Agently-Skills --skill agently-playbook
npx skills add AgentEra/Agently-Skills --skill agently-triggerflow
npx skills add AgentEra/Agently-Skills --skill agently-output-control
npx skills add AgentEra/Agently-Skills --skill agently-model-response
npx skills add AgentEra/Agently-Skills --skill agently-session-memory

migration

bash
npx skills add AgentEra/Agently-Skills --skill agently-playbook
npx skills add AgentEra/Agently-Skills --skill agently-migration-playbook
npx skills add AgentEra/Agently-Skills --skill agently-langchain-to-agently
npx skills add AgentEra/Agently-Skills --skill agently-langgraph-to-triggerflow

6. Recommended project shape ​

The current Skills model does not assume one oversized app.py. It prefers explicit capability boundaries:

  • SETTINGS.yaml or a settings layer for provider config, ${ENV.xxx}, and runtime knobs
  • an app or integration layer that loads settings, validates env names when needed, and calls Agently.set_settings(..., auto_load_env=True)
  • prompts/ for YAML or JSON prompt contracts
  • workflow/ for TriggerFlow graphs and chunk implementations
  • tools/ for replaceable search, browse, MCP, or external adapters
  • outputs/ and logs/ for runtime artifacts instead of mixing them into source folders

If a new request still has an unclear owner layer, project initialization and structure routing should start with agently-playbook.

7. Repo guidance ​

  • The old agent_docs.zip flow is deprecated and no longer the official distribution path.
  • The old local demo skills no longer represent the official skill system.
  • This page stays as the introduction and navigation entry for official Agently Skills.
  • Skill names, installation flows, and catalog structure should follow the current official repository.

For the human-developer view of Agently capabilities and runnable scenarios, continue with: Agent Systems Playbook.