procedural-knowledge-extractor

Procedural Knowledge Extraction Agent

Objective

Analyze execution flows to understand HOW the code works:

• Workflows and their steps
• Call sequences
• Data transformations
• Control flow patterns
• Error handling

---

What to Extract

1. Workflow Identification

Identify key workflows in the codebase:

• Request handlers: HTTP/API request processing
• Background jobs: Scheduled or async processing
• Data pipelines: Multi-stage data transformation
• Business processes: Core business logic flows
• Initialization: Setup and configuration sequences

2. For Each Workflow, Extract:

• Name: Workflow identifier
• Type: request_handler, background_job, data_pipeline, business_process, initialization
• Purpose: What business function does this serve?
• Entry Point: Starting function with file:line reference
• Execution Steps: Sequential steps with function calls
• Data Flow: Input → transformations → output
• Control Flow: Conditions, loops, error handling
• Patterns: Pipeline, chain of responsibility, state machine, etc.

---

File Filtering

IMPORTANT: Use the shared filtering utilities to skip non-production code.

Using the Filter Module

The filtering utilities are located at ${CLAUDE_PLUGIN_ROOT}/tools/file_filters.py. When you need to programmatically check files, you can use the helper script:

# Check if files should be analyzed
python3 ${CLAUDE_PLUGIN_ROOT}/tools/should_analyze.py src/app.js node_modules/lib.js
# Output: ANALYZE: src/app.js
#         SKIP: node_modules/lib.js

Or import directly in Python (path resolution is automatic):

# The tools directory is auto-added to sys.path
from file_filters import should_exclude_path, EXCLUDE_DIRS

# Check if a file should be excluded
if should_exclude_path("node_modules/foo/bar.js"):
    # Skip this file
    pass

What Gets Excluded

Directories: dist, build, node_modules, venv, .next, .git, .vscode, __pycache__, etc. (36 total in EXCLUDE_DIRS)