Skip to content

Use cases

Attestack helps teams answer what happened, what was verified, and whether the record was tampered with — without requiring a SaaS account or always-on surveillance.

You use Cursor, Claude Code, Copilot, or other agents to implement features. Attestack records:

  • Commands you ran (attestack run)
  • Human review notes (attestack note)
  • Git state at start and end (snapshot)
  • Agent tool calls and decisions (attestack agent …)

When you finish, export a signed bundle and attach it to a PR or ticket.

Outcome: Reviewers see structured evidence instead of reconstructing work from chat logs.

A developer attaches an .attestack.zip bundle to a pull request. You run:

Terminal window
attestack verify path/to/bundle.attestack.zip
attestack report --output review.md

Outcome: You verify the event chain and signature offline, then read a human-friendly report.

Security teams need proof that tests ran and changes were reviewed before merge. Attestack captures:

  • Test command exit codes
  • Git HEAD and dirty-state hashes
  • Optional bundle export with path redaction

Outcome: Portable, tamper-evident artifacts suitable for audit trails. Attestack does not claim code correctness — it records what happened.

GitHub Actions (or other CI) runs tests inside an Attestack session, exports a bundle, and uploads it as a workflow artifact. See CI integration.

Outcome: Reproducible CI evidence linked to workflow run IDs.

After a production incident, reconstruct what commands and snapshots were captured during a hotfix session.

Outcome: Timeline with hash chain integrity; tampering is detectable.

Demonstrate reproducible contributor workflows: init → session → tests → bundle → verify.

Outcome: Transparent evidence for security-conscious adopters evaluating the project.

  • Not a replacement for code review or static analysis
  • Not automatic capture of private AI transcripts (opt-in only)
  • Not proof that generated code is secure or correct