mcpField
A
agently-agent-extensions

agently-agent-extensions

github.com/agenteraAgent workflows3141 installsUpdated 4d ago
Repository

About agently-agent-extensions

agently-agent-extensions is a skill for choosing and implementing Agently agent-side extensions. It helps developers prefer native extension surfaces and avoid unnecessary custom dispatchers, wrappers, and bridges.

Key features

  1. 01Selection guidance for native extension surfaces
  2. 02Explicit choices among tools, MCP, and other options
  3. 03Optional agently-devtools installation guidance
  4. 04App-layer wiring through Agently.event_center
  5. 05Advice on built-in browser and desktop support

Use cases

  1. 01Choosing an extension approach for an Agently agent
  2. 02Adding observation or evaluation bridge wiring
  3. 03Implementing browser or desktop control without a custom wrapper

MCP support

The skill identifies MCP as one of Agently's native extension options and advises checking its support before building a parallel tool dispatcher. It does not describe a specific MCP server or protocol configuration.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/agentera/agently-skills --skill agently-agent-extensions

After installing, type/agently-agent-extensionsin Claude Code to confirm it loaded.

Bundled files

References and scripts installed along with SKILL.md.

examples/
.gitkeep
outputs/
.gitkeep
references/
devtools.md
overview.md
scripts/
.gitkeep

Details

From the project README

Agently Agent Extensions

Use this skill when the problem is agent-side extension rather than prompt shape, output contract, or workflow control.

Native-First Rules

  • prefer built-in extension surfaces before handwritten wrappers
  • keep extension choice explicit: tools, MCP, FastAPIHelper, auto_func, KeyWaiter, or agently-devtools
  • treat agently-devtools as an optional companion package installed from PyPI, not as a required source checkout
  • keep observation or evaluation bridge wiring in the app layer through Agently.event_center
  • combine with agently-model-response or agently-triggerflow only when the scenario needs those layers
  • prefer built-in Browse support with Playwright or PyAutoGUI before writing browser or desktop-driving wrappers from scratch

Anti-Patterns

  • do not build a parallel tool dispatcher before checking native tool and MCP support
  • do not create a custom waiter or auto-function shim first
  • do not ask users to clone or editable-install DevTools when pip install agently-devtools is the supported public path
  • do not build a custom runtime upload bridge before checking ObservationBridge

Read Next

  • references/overview.md
  • references/devtools.md

Comments

Markdown supported