Store and retrieve structured JSON variables for SASE agent communication.
npx skills add https://github.com/bbugyi200/dotfiles --skill sase-var
Please help me install this Agent Skill. Skill: sase_var Source: https://github.com/bbugyi200/dotfiles/tree/main/home/dot_codex/skills/sase_var Command: npx skills add https://github.com/bbugyi200/dotfiles --skill sase-var
This skill solves the challenge of passing small, structured data between independent SASE agents, ensuring that downstream processes can reliably consume outputs from upstream tasks.
Use the sase_var skill to set a summary variable with the value tests passed for the current agent run.
High-intent search queries and answers about installing and using this skill.
To pass JSON variables between agents in a SASE workflow, you can set and retrieve structured data types like strings, numbers, lists, and maps within the SASE execution environment. This ensures downstream agents reliably consume upstream outputs.
To conditionally terminate a multi-stage agent chain, you use the reserved STOP variable to signal workflow termination. This mechanism saves compute resources by preventing unnecessary repeat chain executions.
Yes, managing agent output variables for SASE workflows requires a valid SASE_AGENT environment to be present. This environment provides the necessary execution context for storing and retrieving the structured JSON data.
Yes, you can use Jinja templating to consume downstream variables in SASE workflows. The system provides direct integration with Jinja, allowing downstream processes to dynamically render and utilize the structured JSON outputs.
The best way to share build artifacts between independent agents is setting a structured JSON variable, such as a result_path. Downstream agents then retrieve this variable to locate and process the generated artifact.
antgroupAccess Firecrawl API for web search, scraping, and site crawling.
overcut-aiReference Handlebars templating syntax for Overcut workflow data integration.
SalmanferozkhanBuild multi-agent AI workflows in Python with tools, handoffs, and guardrails.
hiddink-aiFacilitate inter-process communication between AI coding agents via a local MCP broker.
