[RFC] Reproducibility Metadata: Auto-generate STUDY.json in Generated Artifacts

[GaneshPatil7517]

Problem

When concore init, concore setup, and concore compose generate study
artifacts (concore.toml, docker-compose.yml, runner.py), they don't record
how and when they were generated. Five years from now, a researcher
reproducing a historical study won't know:

• What tool versions were used when this study was generated?
• Which compiler/runtime was detected on the original machine?
• What concore version created these files?
• Were any manual edits made after generation?

This violates concore's core mission: reproducible science.

Proposed Solution

When concore init --interactive, concore setup, or concore compose
generate a study, automatically create a STUDY.json metadata file:

{
  "generated_by": "concore setup",
  "concore_version": "1.0.0",
  "timestamp": "2026-03-15T10:30:00Z",
  "python_version": "3.13.12",
  "tools_detected": {
    "g++": "6.3.0 at /usr/bin/g++",
    "python3": "3.13.12 at /usr/bin/python3",
    "docker": "not found"
  },
  "platform": "Windows-10",
  "study_name": "my_control_study",
  "checksums": {
    "concore.toml": "sha256:abc123...",
    "docker-compose.yml": "sha256:def456..."
  }
}

Researchers can then verify that the study is reproducible by checking:

1. Tool versions match what was originally available
2. File integrity (checksums haven't drifted)
3. concore version compatibility

Benefits:

1. Generated artifacts become self-documenting
2. Reproducibility is auditable
3. Historical studies can be traced and verified
4. Fits the principle: "Generated files should be independent of the generator"

[GaneshPatil7517]

hello @pradeeban if you agree should i go with this.....?

[pradeeban]

@GaneshPatil7517 You can go ahead and create a PR.