Assemble a Context Envelope and spawn a Task tool subagent to execute tasks.
npx skills add https://github.com/ste-bah/archon --skill run-agent-ste-bah
Please help me install this Agent Skill. Skill: run-agent Source: https://github.com/ste-bah/archon/tree/main/.claude/agents/.claude/skills/run-agent Command: npx skills add https://github.com/ste-bah/archon --skill run-agent-ste-bah
This Skill streamlines the process of launching and coordinating custom agents by automatically assembling the necessary context and delegating task execution to a dedicated subagent.
Provide an agent name and a task description to run, e.g., /run-agent sec-filing-analyzer 'Analyze the latest 10-K filing for insights'.
High-intent search queries and answers about installing and using this skill.
To launch a custom agent, provide an agent name and a task description to automatically assemble a context envelope and spawn a subagent for execution. The tool validates the target agent exists and has a valid agent.md before running.
Context envelope orchestration aggregates agent definitions, memory from MemoryGraph, and LEANN results to form a complete execution context. It injects memory-driven behavioral rules and optional code search context into the task prompt before spawning the subagent.
Running a subagent task requires the agent_name and task arguments, with optional model overrides. The system validates these inputs, ensures the target agent exists, and handles errors before execution.
Yes, you can provide optional model overrides when launching a subagent task. The system accepts these overrides alongside the required agent_name and task arguments to configure the execution.
The structured prompt includes ROLE, DOMAIN CONTEXT, TOOL INSTRUCTIONS, BEHAVIORAL RULES, MEMORY CONTEXT, and YOUR TASK sections. These sections are assembled from the agent definitions and memory context.
Yes, the tool manages tracing and memory insertion throughout the run for later analysis. This ensures subagent execution is auditable and memory context is preserved after the task completes.
AI-Driven-Highspeed-DevelopmentDefine YAML frontmatter and sections for ADHD Framework .agent.md files.
bketelsenAudits memory system architecture and proposes rule changes without inspecting content.
scooter-lacroixCompare code search tools and select via decision tree.
DNYoussefCoordinate multi-agent GitHub pull request reviews with parallel security, performance, and code analysis.
