Artifacts Reference

prompts-gpt generates artifact files for every execution. This reference documents the artifact directory structure and file formats.

Artifacts Directory

Default location: .scripts/runs/

Configurable via:

Config: "artifactsDir": ".scripts/runs"
CLI: --artifacts-dir path/to/dir
Env: PROMPTS_GPT_ARTIFACTS_DIR

Directory Structure

Single Run

.scripts/runs/
└── agent-{timestamp}-{hash}/
    ├── output.log        # Raw agent CLI output
    ├── summary.md        # Extracted summary
    └── meta.json         # Run metadata

Batch Run

.scripts/runs/
└── batch-{timestamp}-{hash}/
    ├── prompt-1/
    │   ├── output.log
    │   ├── summary.md
    │   └── meta.json
    ├── prompt-2/
    │   ├── output.log
    │   ├── summary.md
    │   └── meta.json
    └── batch-summary.json

Sweep

.scripts/runs/
└── sweep-{timestamp}-{hash}/
    ├── iteration-1/
    │   ├── output.log
    │   ├── summary.md
    │   └── meta.json
    ├── iteration-2/
    │   ├── output.log
    │   ├── summary.md
    │   └── meta.json
    ├── iteration-3/
    │   ├── output.log
    │   ├── summary.md
    │   └── meta.json
    ├── sweep-manifest.json
    └── eval-scores.json      # If eval enabled

Orchestration

.scripts/runs/
└── orchestrate-{timestamp}-{hash}/
    ├── provider-codex/       # Parallel mode
    │   ├── output.log
    │   └── summary.md
    ├── provider-claude/
    │   ├── output.log
    │   └── summary.md
    └── orchestrate-summary.json

File Formats

meta.json

{
  "runId": "agent-20260521-abc123",
  "provider": "codex",
  "model": "gpt-5.5",
  "promptFile": ".prompts-gpt/review.md",
  "startedAt": "2026-05-21T14:30:00Z",
  "completedAt": "2026-05-21T14:32:15Z",
  "durationMs": 135000,
  "exitCode": 0,
  "tokenUsage": {
    "inputTokens": 1250,
    "outputTokens": 3400,
    "totalTokens": 4650
  },
  "gitBranch": "main",
  "gitHead": "abc1234"
}

sweep-manifest.json

{
  "sweepId": "sweep-20260521-abc123",
  "promptFile": ".prompts-gpt/sweeps/audit.md",
  "provider": "codex",
  "model": "gpt-5.5",
  "requestedIterations": 5,
  "completedIterations": 5,
  "startedAt": "2026-05-21T14:30:00Z",
  "completedAt": "2026-05-21T15:00:00Z",
  "totalDurationMs": 1800000,
  "iterations": [
    {
      "iteration": 1,
      "status": "completed",
      "durationMs": 320000,
      "exitCode": 0
    }
  ]
}

eval-scores.json

{
  "scores": [
    {
      "iteration": 1,
      "overallScore": 0.72,
      "criteria": [
        { "name": "correctness", "score": 0.8 },
        { "name": "completeness", "score": 0.65 }
      ],
      "passed": true
    }
  ],
  "passThreshold": 0.7
}

Artifact Pruning

Old artifact directories are automatically pruned. Controlled by:

{
  "maxRunDirs": 20
}

Or via CLI: --max-run-dirs 10

Viewing Artifacts

# Show latest run directory
prompts-gpt diff

# List artifact directories
ls .scripts/runs/

# View specific summary
cat .scripts/runs/agent-20260521-abc123/summary.md