Skill: Template Candidates

Command: /template-candidates [--category <type>] Model: 🟡 Sonnet Agent: None (direct execution)

Purpose

Analyse files in this vault and identify candidates suitable for the ArchitectKB template repository. Performs detailed content analysis to ensure files are generic and reusable.

When to Use

• After /template-changelog identifies new/modified files
• Before running /template-sync
• When deciding what features to contribute upstream

Categories

Filter by category with --category:

Category	Directory	Description
scripts	.claude/scripts/	Python/Shell automation
skills	.claude/skills/	Skill definitions
schemas	.claude/schemas/	JSON validation schemas
hooks	.claude/hooks/	Pre/post execution hooks
rules	.claude/rules/	Reference documentation
templates	Templates/	Note templates
all	All above	Full analysis

Implementation

Step 1: Scan for New/Modified Files

Compare with template repo:

VAULT = Path("~/Documents/GitHub/BA-DavidOliver-ObsidianVault").expanduser()
TEMPLATE = Path("~/Documents/GitHub/ArchitectKB").expanduser()

DIRS_TO_COMPARE = [
    ".claude/scripts",
    ".claude/skills",
    ".claude/schemas",
    ".claude/hooks",
    ".claude/rules",
    "Templates",
]

Step 2: Content Analysis

For each candidate file, check for:

Disqualifying Content (AUTO-EXCLUDE)

Pattern	Reason
Acme Corp, BA-, britishairways	Company-specific
@ba.com, atlassian.net/wiki/spaces/BA	Internal URLs
API keys, tokens, passwords	Security risk
/Users/david.oliver/	Hardcoded personal paths
ERPSystem, Alpha, Beta, DataPlatform, AlertHub	BA-specific projects/systems
notion.so/ba-workspace	Internal integrations

Qualifying Content (GOOD SIGNS)

Pattern	Reason
Generic variable names	$VAULT_PATH, ${PROJECT}
Configuration via environment	os.environ.get()
Template placeholders	{{company}}, <your-org>
Well-documented functions	Docstrings, comments
Error handling	Try/except, validation

Step 3: Scoring System

Score each file 0-100:

Score	Classification	Action
90-100	Ready	Can sync directly
70-89	Minor edits	Small sanitisation needed
50-69	Needs work	Significant generalisation required
0-49	Not suitable	Too specific or contains secrets

Step 4: Generate Report

Output a structured report:

# Template Candidates Report

Generated: 2026-02-01

## Summary

| Category | Ready | Minor Edits | Needs Work | Excluded |
|----------|-------|-------------|------------|----------|
| scripts  | 3     | 2           | 1          | 4        |
| skills   | 5     | 3           | 2          | 1        |
| ...      |       |             |            |          |

## Ready for Sync (Score 90+)

### .claude/scripts/archive-orphan-attachments.py
- **Score:** 95
- **Status:** Ready
- **Notes:** Fully generic, no company references

### .claude/schemas/location.schema.json
- **Score:** 100
- **Status:** Ready
- **Notes:** Pure schema, no customisation needed

## Minor Edits Needed (Score 70-89)

### .claude/skills/sync-notion/SKILL.md
- **Score:** 75
- **Status:** Minor edits
- **Issues:**
  - Line 23: Contains hardcoded database ID
  - Line 45: References BA-specific workspace
- **Fix:** Replace with placeholder variables

## Excluded (Score <50)

### .claude/scripts/sync-ba-confluence.py
- **Score:** 20
- **Reason:** Entirely BA-specific integration
- **Action:** Do not sync

## Recommended Actions

1. Run `/template-sync` for 8 ready files
2. Edit 5 files with minor issues
3. Review 3 files needing significant work

Output Location

Save report to: .claude/reports/template-candidates-YYYY-MM-DD.md

Related Skills

• /template-changelog - Generate changelog entries
• /template-sync - Execute the sync