agently-input-composition

Installation
SKILL.md

Agently Input Composition

This skill covers how to compose model input in Agently before the request is sent. It focuses on prompt slots, prompt layering, quick prompt methods, placeholder mappings, serializable prompt data, low-level chat_history, and attachments. It does not cover model setup, output schema control, YAML/JSON prompt template files, or session lifecycle management.

Prerequisite: Agently >= 4.0.8.5.

Scope

Use this skill for:

  • choosing between agent-level and request-level prompt state
  • using set_agent_prompt(...) and set_request_prompt(...)
  • using quick prompt methods such as system(), role(), rule(), user_info(), input(), info(), instruct(), examples(), and attachment()
  • deciding when to use always=True
  • composing input with standard prompt slots
  • using placeholder mappings in prompt keys and values
  • passing lists, dicts, and other serializable data as prompt content
  • using low-level chat_history as input-side context
  • using attachment() for rich-content input
Related skills

More from agentera/agently-skills

Installs
1
GitHub Stars
8
First Seen
Mar 11, 2026