Security Review Agent Skill
Threat modeling checklist for web apps: auth, secrets, injection, and abuse cases.
Connected Apps: GitHub Claude
Installation Guide
### Use this Agent Skill
1. Open your agent client (Claude, Cursor, or custom agent loop).
2. Paste the configuration JSON into your skills / system prompt library.
3. Invoke the skill with a concrete goal and any required context files.
4. Review outputs against the checklist in the skill body before shipping.
Detailed Documentation
What this skill does
Threat modeling checklist for web apps: auth, secrets, injection, and abuse cases.
When to use it
- You want a repeatable agent behavior instead of ad-hoc prompting
- You need structured outputs (lists, tables, JSON, checklists)
- You are chaining this skill with MCP tools or automation workflows
Checklist
- State the goal in one sentence
- Provide constraints (tone, length, tools allowed)
- Attach relevant context (files, URLs, CRM fields)
- Require citations or verification steps when facts matter
- End with owners / next actions
Pair with
- MCP Directory for tool access
- Workflows for multi-step automation
- Tools for platform-specific recipes
Attribution
Editorial agent skill template by WorkflowAI. Customize prompts for your stack before production use.
Configuration
{
"agentSkills": {
"agent-skill-security-review": {
"name": "Security Review Agent Skill",
"description": "Threat modeling checklist for web apps: auth, secrets, injection, and abuse cases.",
"prompt": "You are the Security Review Agent Skill. Follow the skill checklist: clarify goal, gather context, produce structured output, list assumptions and next steps."
}
}
}