mcg-mentioned-coverage-gate

name	mcg - Mentioned Coverage Gate
description	Gate that verifies all items mentioned by user have been considered, analyzed, and addressed. Prevents dropping user-mentioned aspects. Run after any multi-aspect analysis to verify nothing was missed.
output	{"format":"prose"}

Mentioned Coverage Gate

Input: $ARGUMENTS (the original user message or list of items to verify coverage for)

---

Purpose

When a user mentions multiple things (frames, stakeholders, dimensions, categories, etc.), this gate verifies that EVERY mentioned item was:

1. Identified/extracted
2. Checked against existing coverage
3. Either covered or explicitly noted as a gap
4. Gaps filled or explained

Why this gate exists: It's easy to address some items and unconsciously drop others. This gate prevents that.

---

Step 1: EXTRACT All Mentioned Items

Parse the user's input and create an exhaustive list of everything they mentioned.

Categories to look for:

• Explicit lists (numbered, bulleted, comma-separated)
• Items in tables
• Items in examples
• Implicit items (references to "all of these", "each of", "etc.")
• Meta-items ("also" statements, "and" connectors)

Output format:

EXTRACTED ITEMS:
1. [item 1] - from: [where in message]
2. [item 2] - from: [where in message]
...
N. [item N] - from: [where in message]

Total: N items extracted

Gate check: Did I extract ALL items, including:

• Items explicitly listed?
• Items implied by "etc." or "and so on"?
• Items referenced with "all of these"?
• Items mentioned in passing?
• Meta-requests about the items themselves?

---

Step 2: CHECK Coverage for Each Item

For each extracted item, determine if it's covered in the system.

Check locations:

1. Universal guess library files (data/guess_libraries/universal/*.md)
2. Domain guess library files (data/guess_libraries/*.md)
3. Skills (skills/*/SKILL.md)
4. Procedures in YAML files

Output format:

| Item | Covered? | Location | Specific Question/Entry |
|------|----------|----------|------------------------|
| [item 1] | YES/NO/PARTIAL | [file] | [Q# or entry] |
| [item 2] | YES/NO/PARTIAL | [file] | [Q# or entry] |
...

Coverage definitions:

• YES: Explicit question-guess entry exists for this exact concept
• PARTIAL: Related content exists but doesn't fully address the item
• NO: No coverage found

---

Step 3: IDENTIFY Gaps

Create explicit list of items not covered or only partially covered.

Output format:

GAPS IDENTIFIED:

1. [item] - Status: [NO/PARTIAL]
   - What's missing: [specific gap]
   - Why it matters: [impact of not having this]
   - Suggested file: [where it should go]

2. [item] - Status: [NO/PARTIAL]
   ...

Total gaps: N items

---

Step 4: FILL or EXPLAIN Gaps

For each gap, either:

1. CREATE the missing content (new file or new questions in existing file)
2. EXPLAIN why it shouldn't be created (out of scope, duplicate, etc.)

If creating:

CREATED:
- [item] → [new file or location]
  - Questions added: [count]
  - Entries added: [count]

If explaining:

NOT CREATED (with reason):
- [item] → Reason: [why not needed]

---

Step 5: VERIFY Complete Coverage

After filling gaps, run coverage check again to confirm 100% coverage.

Final output format:

FINAL COVERAGE VERIFICATION:

| Item | Status | Location |
|------|--------|----------|
| [item 1] | ✓ COVERED | [file:Q#] |
| [item 2] | ✓ COVERED | [file:Q#] |
...

Coverage: N/N items (100%)

Gate FAILS if:

• Any item shows "NOT COVERED" without explanation
• Any item was dropped (not in final table)
• Total in final table ≠ Total extracted in Step 1

---

Step 6: META-CHECK

Verify the gate itself was applied correctly:

• Did I extract ALL items from user message?
• Did I check EVERY extracted item?
• Did I create or explain EVERY gap?
• Does final count match initial count?
• Did user mention any META-requests about the items that I also need to address?

---

Quick Reference Table

Step	Action	Output
1	Extract	List of N items
2	Check	Coverage table
3	Identify	Gap list
4	Fill/Explain	Created/Explained list
5	Verify	Final 100% coverage table
6	Meta-check	Self-verification

---

Example Application

User says: "Make sure frames, stakeholders, time horizons, and dimensions are all covered"

Step 1 extraction:

EXTRACTED ITEMS:
1. frames - explicit
2. stakeholders - explicit
3. time horizons - explicit
4. dimensions - explicit
5. "all covered" - meta-request for verification

Total: 5 items (4 content + 1 meta)

Step 2 check:

| Item | Covered? | Location |
|------|----------|----------|
| frames | YES | 104_interpretation_frames.md |
| stakeholders | YES | 105_stakeholder_archetypes.md |
| time horizons | YES | 106_time_horizons.md |
| dimensions | PARTIAL | 01-11 exist but DEGREE missing |
| "all covered" meta | NO | No coverage verification file |

Step 3 gaps:

1. dimensions/DEGREE - PARTIAL → Need 107_degree_magnitude.md
2. "all covered" meta → Need this very procedure

Step 4 fill:

CREATED:
- DEGREE → 107_degree_magnitude.md (11Q, 51 entries)
- coverage gate → skills/mcg/SKILL.md

Step 5 verify:

Coverage: 5/5 items (100%)

---

When To Use This Gate

ALWAYS use after:

• User lists multiple items to address
• User says "all of these", "each of", "make sure"
• Complex multi-aspect analysis
• Creating new categories or files

SIGNS you need this gate:

• User repeats a request (may have been dropped)
• User says "did you consider X?" (may have been missed)
• You feel like you might have missed something

---

Integration

This gate should be invoked:

1. Automatically after /spd (verify all discovered aspects addressed)
2. Manually when user mentions multiple items
3. As final check before declaring work complete

→ Pairs with: /comprehensive_aspects (general completeness) → Pairs with: /vbo (claim verification) → Pairs with: 100_coverage_strategy.md (coverage modes)