Slash Command

/detect-patterns

Detect repetitive workflow patterns from Claude Code session history and propose automations.

Install

npx claudepluginhub ai-sdlc-framework/ai-sdlc --plugin ai-sdlc

Referenced Files

ai-sdlc/orchestrator

Prompt Preview

Detect repetitive workflow patterns from Claude Code session history and propose automations.

## Steps

1. **Read telemetry data** from `~/.claude/usage-data/tool-sequences.jsonl`
2. **Also read** session metadata from `~/.claude/usage-data/session-meta/*.json` for historical data
3. **Run pattern detection** using the orchestrator's n-gram mining engine:
   
4. **Classify each pattern** as command-sequence, copy-paste-cycle, or periodic-task
5. **Generate proposals** for each detected pattern with draft artifact content
6. **Present results** as a table:
   | # | Pattern | Type | Confiden...

Command Content

Other plugins with /detect-patterns

/detect-patterns

history-analysis/

**Purpose:** Identify project-specific commit message patterns, conventions, and formatting standards from commit history.

git-commit-assistant

Stats

Parent Repo Stars8

Parent Repo Forks2

Last CommitApr 2, 2026

Actions

View Source View Plugin View on GitHub View README

Detect repetitive workflow patterns from Claude Code session history and propose automations. ## Steps 1. **Read telemetry data** from `~/.claude/usage-data/tool-sequences.jsonl` 2. **Also read** session metadata from `~/.claude/usage-data/session-meta/*.json` for historical data 3. **Run pattern detection** using the orchestrator's n-gram mining engine: 4. **Classify each pattern** as command-sequence, copy-paste-cycle, or periodic-task 5. **Generate proposals** for each detected pattern with draft artifact content 6. **Present results** as a table: | # | Pattern | Type | Confiden...

Detect repetitive workflow patterns from Claude Code session history and propose automations.

Steps

Read telemetry data from ~/.claude/usage-data/tool-sequences.jsonl
Also read session metadata from ~/.claude/usage-data/session-meta/*.json for historical data

Run pattern detection using the orchestrator's n-gram mining engine:

import { readToolSequenceJSONL, mineFrequentPatterns, classifyPattern } from '@ai-sdlc/orchestrator';

Classify each pattern as command-sequence, copy-paste-cycle, or periodic-task
Generate proposals for each detected pattern with draft artifact content
Present results as a table: | # | Pattern | Type | Confidence | Frequency | Sessions | Proposed Artifact |
For each proposal, show the draft artifact content and ask if the user wants to approve it
If approved, write the artifact file using the artifact writer

Arguments

--since YYYY-MM-DD — only analyze events after this date
--min-confidence 0.6 — minimum confidence threshold (0-1)

Important

Only create artifacts the user explicitly approves
Never overwrite existing files
Show the full draft content before writing