LLAJinja
github.com/electronick1/llajinja- Category
- AI Agents
- Rank
- No. 3687Tools index
Previous survey · No. 3549 ·
- Pricing
- Open Source
- Type
- TOOL
- Use case
- Agent Building
- Interfaces
- Agent Skill / Plugin
- GitHub
- 11 stars
- Date
About
LLAJinja is a Skill that lets AI coding harnesses such as Claude Code, OpenCode, and Codex run other Skills written as Jinja templates, without requiring extensions or plugins. It adds templating features including declared input variables that prompt the user when missing, sequential stages for multi-step skills, and agent blocks that spawn sub-agents with defined input and output schemas.
At a glance
LLAJinja describes agent skills as templates with inputs, stages and outputs that later steps can use. It is worth evaluating when the same procedure keeps getting rewritten as a long prompt.
Good fit
People maintaining a recurring workflow in an agent harness that can load skills.
Consider something else when
Use an ordinary checklist for one-off tasks or an execution engine when steps need hard enforcement.
Start here
Write two harmless stages where the second uses a value from the first. Leave one input blank and check whether the harness asks for it and carries the answer forward.
Cost and requirements
Work-in-progress skill for compatible harnesses. Delegated stages require the harness to support subagents.
Limits to know
Behavior depends on the host’s capabilities. Treat templates as instructions and inspect any nested skill before using it.
Sources checked 2026-11-05. Source-based assessment.
Tags
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.