engagement merge

Combine multiple engagement JSON or JSONL files into a single output for scoring.

Synopsis

aipostex engagement merge [file1.json] [file2.json] ... [flags]

Description

The engagement merge command reads multiple aipostex engagement JSON documents or JSONL finding streams and combines them into a single unified engagement document. Duplicate findings are removed based on a content hash of the target, title, and evidence fields.

This is useful when running separate module commands (e.g., ollama prompts, vectordb search-sensitive, ray jobs) and needing to combine results into one file for scoring or reporting.

Flags

Flag	Description
--output, -o	Output file path (default: stdout)

Only --output applies to this command.

Deduplication

Findings are deduplicated by SHA-256 hash of target + title + evidence. When duplicates exist, the earliest finding (by timestamp) is kept.

Examples

# Merge three result files into one
aipostex engagement merge ollama.json vectordb.json ray.json -o combined.json

# Merge all JSON files in a directory
aipostex engagement merge results/*.json -o assessment.json

# Merge and output to stdout
aipostex engagement merge scan-results.json mlflow-results.json

Workflow

Run individual module commands with --format json or --format jsonl --output <file>
  → aipostex engagement merge <file1> <file2> ... -o combined.json
    → aipostex report summary combined.json