name: dedup disable-model-invocation: true description: | Detect semantic code duplication — functions that do the same thing but have different names or implementations. Use when asked to "find duplicates", "check for duplicate functions", "consolidate utilities", or before major refactoring efforts. license: MIT metadata: author: howells website: order: 21 desc: Semantic duplicate finder summary: Finds functions that do the same thing but have different names or implementations. what: | Dedup scans your codebase for semantic duplicates — functions that serve the same purpose with different names or implementations. It extracts a function catalog, categorizes by domain, identifies consolidation opportunities with confidence ratings, and presents a prioritized list of what to merge. why: | Codebases accumulate duplicates organically. Three different formatDate functions, five string sanitizers, two path normalizers. This command finds them so you can consolidate before the codebase becomes unmaintainable.

<progress_context> Use Read tool: docs/progress.md (first 50 lines)

Check for recent refactoring context that might inform duplicate detection scope. </progress_context>

Duplicate Detection Workflow

Name: dedup
Rating: 78
Author: howells

Run the 5-phase duplicate detection process using the duplicate-detector research agent.

Process

•Spawn the duplicate-detector agent with the current codebase scope
•Present consolidated results — grouped by domain with confidence ratings
•Offer consolidation plan — suggest which duplicates to merge and into what

Execution

Spawn the duplicate-detector agent:

code

Agent: ${CLAUDE_PLUGIN_ROOT}/agents/research/duplicate-detector.md

Pass the codebase scope (default: full codebase, or user-specified path).

After Detection

Present results in this format:

markdown

## Duplicate Detection Results

### High Confidence (≥90%)
| Function A | Function B | Domain | Recommendation |
|-----------|-----------|--------|---------------|
| formatDate() in utils/date.ts | formatTimestamp() in helpers/time.ts | Date formatting | Consolidate into formatDate() |

### Medium Confidence (70-89%)
[Same table format]

### Low Confidence (50-69%)
[Same table format — mention these but don't recommend action]

Ask the user which duplicates they want to consolidate, then offer to create a plan.

<progress_append> After running duplicate detection, append to progress journal:

markdown

## YYYY-MM-DD HH:MM — /arc:dedup
**Task:** Duplicate detection
**Outcome:** Found X duplicate groups (Y high confidence, Z medium)
**Scope:** [path or "full codebase"]
**Next:** Consolidate [N] duplicate groups

---

</progress_append>