MANDATORY workflow for creating new agents. Enforces -01-agent-formatting standards from inception, ensuring every new agent is properly designed, researched, and curated before deployment.
SkillScore83Maintained
Automatic
◆AI Engineer pack
Last verified 8/18/2026
What it changes
Passed 17 of 21 test cases with the skill, and 5 of 21 without it.
Without skill22%
With skill83%
Checks it passed that the model alone missed
The agent is assigned to PhD tier.
The agent is assigned to Expert tier.
The agent is assigned to Focused tier.
and 21 more in the full report
16 fixed3 passed either way2 still failing2 broke
Pass-rate lift
+61 pp
Tokens
+201%
Agent turns
0%
Response time
-58%
Model gemini-3.6-flashMeasured 8/18/2026v119 of 23 cases passed with skill · 21 comparable
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 15 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 15
Prompt
During Step 7 Synthesis & Validation, what save option is provided in the menu frame?
What was checked, without and with the skill
Checked
Without skill
With skill
The menu frame contains option [S] Save agent.
Missed
Passed
case-18 · gemini-3.6-flash · 8/18/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md397 linessingle file
Agent Provisioner Skill
Purpose
This is the ONLY sanctioned way to create new agents in dev-system.
Creating agents without this skill results in:
Missing knowledge source validation
Unverified tier classification
Incomplete MCP integration
Poor instruction quality
No curation record
Activation
/agent-provisioner
Parameters:
domain: {domain-description}
name: {proposed-agent-name}
category: {parent-category in -03-agents}
Or simply:
I need a new agent for {description}
Workflow Overview
┌─────────────────────────────────────────────────────────────────────────┐
│ AGENT PROVISIONING WORKFLOW │
├─────────────────────────────────────────────────────────────────────────┤
│ │
│ Step 1: DOMAIN ANALYSIS │
│ ├─ What domain does this agent cover? │
│ ├─ What would a PhD in this domain know? │
│ └─ What existing agents overlap? (avoid duplication) │
│ ▼ │