ring/dev-team/skills/shared-patterns/standards-boundary-enforcement.md

Standards Boundary Enforcement

Canonical source for preventing agents from hallucinating requirements beyond what exists in Ring standards files.

---

⛔ CRITICAL: Check only What Standards Define

This is NON-NEGOTIABLE. Agents MUST not invent, assume, or hallucinate requirements.

Rule	Enforcement
only check items explicitly listed in standards	If not in WebFetch result → not a requirement
never invent "should have" items	Standards define WHAT exists, not what you think should exist
never add industry best practices not in standards	Ring standards ARE the best practices for this org
never assume common patterns are required	If not listed → not required

If you flag something not in standards → Your output is INVALID. Remove the hallucinated finding.

---

Why This Pattern Exists

Problem: AI agents "helpfully" invent requirements based on:

• General industry knowledge
• Common patterns seen in training data
• Assumptions about what "should" exist
• Extrapolation from partial information

Example Failures:

Agent Said	Reality	Problem
"Missing make proto"	not in devops.md Makefile Standards	Hallucinated requirement
"Missing make mocks"	not in devops.md Makefile Standards	Hallucinated requirement
"Missing make migrate-up/down"	not in devops.md Makefile Standards	Hallucinated requirement
"Should use gRPC"	not in golang.md Frameworks	Hallucinated requirement
"Needs GraphQL schema"	not in typescript.md	Hallucinated requirement

Solution: Agents MUST extract requirements only from WebFetch result, never from general knowledge.

---

⛔ MANDATORY: Standards Extraction Process

Before checking compliance, you MUST:

Step 1: WebFetch the Standards File

WebFetch:
  url: [agent-specific URL from standards-workflow.md]
  prompt: "Extract all requirements, patterns, and expected items"

Step 2: Extract and LIST the Requirements

For each section you will check, you MUST output what you found:

## Standards Extracted from WebFetch

### Section: [Name]

**Requirements found in standards file:**
1. [Exact requirement from WebFetch]
2. [Exact requirement from WebFetch]
3. [Exact requirement from WebFetch]

**I will check only these items. Nothing else.**

Step 3: Check only Extracted Requirements

Your findings MUST reference the extracted list:

### Finding: [Issue]

**Standard Reference:** [Section Name] → Requirement #2 from extracted list
**Evidence:** [file:line showing violation]

---

Agent-Specific Boundaries

⛔ CRITICAL: Do not duplicate standards content here. Reference the standards files.

How to Use This Pattern

Each agent MUST:

1. WebFetch the appropriate standards file
2. Extract requirements from the fetched content
3. Check only what the standards explicitly list
4. Verify before flagging - if not in WebFetch result, DO NOT flag

Common Hallucinations by Agent Type

These are items agents commonly invent that are not typically in standards. Always verify in WebFetch result before flagging.

ring:devops-engineer → devops.md

Common Hallucination	Action
make proto	Verify in devops.md "Makefile Standards" section
make mocks	Verify in devops.md "Makefile Standards" section
make migrate-*	Verify in devops.md "Makefile Standards" section
make install	Verify in devops.md "Makefile Standards" section
make clean	Verify in devops.md "Makefile Standards" section

ring:backend-engineer-golang → golang.md

Common Hallucination	Action
gRPC requirement	Verify in golang.md "Frameworks & Libraries" section
GraphQL requirement	Verify in golang.md "Frameworks & Libraries" section
Gin instead of Fiber	Check actual HTTP framework in golang.md
GORM instead of pgx	Check actual ORM/driver in golang.md

ring:backend-engineer-typescript → typescript.md

Common Hallucination	Action
class-validator	Verify validation library in typescript.md
TypeORM	Verify ORM in typescript.md
Jest	Verify testing framework in typescript.md
InversifyJS	Verify DI framework in typescript.md

ring:sre → sre.md

Common Hallucination	Action
Prometheus metrics endpoint	Verify in sre.md if explicitly required
Jaeger UI	Verify in sre.md if explicitly required
Custom dashboards	Verify in sre.md if explicitly required
Alert rules	Verify in sre.md if explicitly required

ring:frontend-designer → frontend.md

Common Hallucination	Action
Design library (Material UI, Chakra)	Verify in frontend.md "Libraries & Tools" section
CSS framework (Bootstrap, Bulma)	Verify in frontend.md "Styling Standards" section
Accessibility standard (WCAG AAA)	Verify EXACT level in frontend.md "Accessibility" section
Component library (Radix, Headless UI)	Verify in frontend.md "Component Patterns" section

---

Anti-Rationalization Table

Rationalization	Why It's WRONG	Required Action
"Industry standard to have make proto"	Industry ≠ Ring standards. Ring defines requirements.	Do not flag
"Most Go projects need gRPC"	Most ≠ this project. Standards define this project.	Do not flag
"It's a best practice to have X"	Best practices are IN the standards. If not there, not required.	Do not flag
"This would improve the codebase"	Improvement suggestions ≠ compliance findings.	Do not flag as non-compliant
"I've seen this in similar projects"	Similar ≠ this. Standards are project-specific.	Do not flag
"Common sense says this is needed"	Common sense ≠ explicit requirement. Standards are explicit.	Do not flag
"The team might need this"	Probably ≠ definitely. Standards define definitely.	Do not flag
"It's implied by the architecture"	Implied ≠ explicit. Only explicit requirements count.	Do not flag

---

Self-Verification Checklist

Before submitting findings, verify:

For each finding:
1. Is this item EXPLICITLY listed in the WebFetch result?       [ ]
2. Can I quote the EXACT line from standards defining this?     [ ]
3. Am I flagging absence of something not in standards?         [ ]
4. Am I inventing requirements from general knowledge?          [ ]
5. Am I assuming "should have" vs "standards say must have"?    [ ]

If #1 or #2 is no → REMOVE the finding
If #3, #4, or #5 is YES → REMOVE the finding

---

How Agents Reference This Pattern

Add to agent's Standards Compliance section:

## Standards Boundary (MANDATORY)

See [shared-patterns/standards-boundary-enforcement.md](../skills/shared-patterns/standards-boundary-enforcement.md) for:
- only check what standards explicitly define
- Agent-specific requirement boundaries
- Forbidden hallucinations list
- Self-verification checklist

**⛔ HARD GATE:** Before flagging any item as non-compliant:
1. Verify the requirement EXISTS in WebFetch result
2. Quote the exact standard that requires it
3. If you cannot quote it → Do not flag it