// Load when writing or reviewing agent .md files. Design guidance for agent definitions and meridian-specific conventions.
Load when writing, reviewing, or improving any prompt artifact. Research-backed principles across four levels: prompt craft, skill extraction, agent design, and multi-agent coordination.
Load when reviewing agent definitions or skill content. Adversarial review methodology — finding quality, severity, and structured reporting.
Load when writing or reviewing skills. Design guidance for SKILL.md files and bundled resources.
| name | agent-artifacts |
| type | reference |
| description | Load when writing or reviewing agent .md files. Design guidance for agent definitions and meridian-specific conventions. |
| model-invocable | false |
Agent definitions are markdown files with YAML frontmatter. For the full frontmatter schema (fields, values, harness overrides, model policies, fanout), see the mars agent profile reference.
The body below frontmatter is the system prompt. Opening matters most — it survives compaction and gets strongest attention. Let skills carry methodology the agent shares with others; the body carries what's unique to this agent.
Descriptions serve callers, not the agent. Lead with the problem it solves and when to reach for it, then teach how to use it effectively.
Managers and leads coordinate — they don't implement. Use
disallowed-tools: [Agent] to force meridian spawn (which provides
tracking and reports).
For meridian-specific conventions (subagent design, context handoffs, model
staffing), see resources/meridian.md.