How this skill is triggered — by the user, by Claude, or both
Slash command
/prd-generator:prd-guardianThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
---
This skill activates in any project generated or managed by prd-generator-plugin. It ensures that AI coding agents always consult project-specific enforcement skills before taking action.
digraph prd_guardian {
"About to implement a feature?" [shape=diamond];
"Making an architectural decision?" [shape=diamond];
"Reviewing or refactoring code?" [shape=diamond];
"Load project skills" [shape=box];
"Proceed" [shape=doublecircle];
"About to implement a feature?" -> "Load project skills" [label="yes"];
"Making an architectural decision?" -> "Load project skills" [label="yes"];
"Reviewing or refactoring code?" -> "Load project skills" [label="yes"];
"About to implement a feature?" -> "Making an architectural decision?" [label="no"];
"Making an architectural decision?" -> "Reviewing or refactoring code?" [label="no"];
"Reviewing or refactoring code?" -> "Proceed" [label="no"];
"Load project skills" -> "Proceed";
}
Before loading project skills, check the version header:
.claude/skills/project-guardian/SKILL.md → prd_version: X.Y
docs/prd/PRD.md → Version: X.Y (in the header)
If versions differ: warn the user — "Project skills are out of sync with PRD. Run /prd-evolve to sync before proceeding."
Do NOT block development if stale — warn and let the user decide. Do flag if the discrepancy seems material to the current task.
These trigger regardless of what the project skills say:
| Situation | Action |
|---|---|
No docs/prd/PRD.md found | Warn — suggest running /prd-new |
No .claude/skills/project-guardian/SKILL.md found | Warn — project skills not generated |
User asks to remove a feature without /prd-evolve | Remind to run /prd-evolve first |
| File | Purpose |
|---|---|
docs/prd/PRD.md | Product requirements source of truth |
docs/prd/ARCHITECTURE.md | Canonical stack and architectural decisions |
docs/prd/ER.md | Entity relationship model |
.claude/skills/project-guardian/ | PRD compliance enforcement |
.claude/skills/project-architecture/ | Technical stack enforcement |
.claude/skills/project-domain-rules/ | Business logic invariants |
.claude/skills/project-compliance/ | Regulatory requirements |
.claude/skills/project-docs/ | Living documentation maintenance |
docs/index.md | Project documentation entry point |
backend/CLAUDE.md | Backend development standards |
frontend/CLAUDE.md | Frontend development standards |
infrastructure/CLAUDE.md | Infrastructure standards |
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
npx claudepluginhub rodrigorjsf/prd-generator-plugin