Explore the portfolio

Open to AI/ML and software engineering internships.

Back to portfolio

AgentSentry v2

An approval-gated support-action system with durable execution, explicit authority, recipient receipts, and recovery after worker failure.

Measured development system
AgentSentry workspace with an approved action, execution state, and audit events
Actual governed-action workspace with synthetic orders and a simulated recipient service.Open full capture

Solo project; AI-assisted development.

Completed synthetic actions

128 / 128

13 October 2026 hosted bounded workload, with 128 recipient effect receipts. Four callers, two workers; not saturation or a production SLA.

Validation notes

The idea

The problem

An action can outlive an HTTP response or a worker. Retrying without recipient evidence can repeat the external effect.

The implementation

What I built

Completed 128 synthetic actions with 128 recipient receipts in a bounded hosted run by combining approval-bound actions, fenced workers, and receipt reconciliation.

  • Go
  • PostgreSQL
  • Python
  • JavaScript
  • Docker

Features

  • Propose and approve exact action hashes
  • Inspect execution state and audit events
  • Reconcile recipient receipts after failure

Architecture

  • Go API, PostgreSQL approval/action storage, and a separate Go worker.
  • Browser workspace for proposals, approval, saved state, and audit events.
  • Authenticated simulated order API with SQLite receipts and an experimental typed model-draft path.

Engineering decisions

  • Bind approval to the exact action hash; strict mode requires a different authorized approving actor.
  • Check recipient receipts before retrying an uncertain action.
  • Separate model draft quality from execution correctness.

Validation and measured results

  • 13 October 2026 hosted CI verified a dated engineering baseline, including recovery, race checks, browser workflows, and restore drills.
  • A bounded synthetic hosted workload completed 128 actions with 128 effect receipts; it is not a production SLA or saturation result.
  • The frozen model evaluation stopped at 62/80 cases and already failed its declared utility and safety gates.

Limits and unfinished work

  • AI quality qualification remains open. Engineering checks do not qualify autonomous decisions.
  • The simulated recipient supplies explicit idempotency and receipt guarantees that arbitrary external APIs may lack.
  • No customer deployment or API cost saving is claimed.

Evidence and sources

  • Dated engineering validation notes

    13 October 2026 hosted run 38379352933 at commit 37517e29913e6ec66bb4cce6a4c5373785eb3370 was inspected during this review. Repository and CI links were not publicly accessible on 14 October; no downloadable receipt is attached.

Explore the project archive

Find your way

12 results found.

Use Tab or arrow keys to explore. Enter opens a result. Escape closes search.

Portfolio assistant

Answers from published portfolio records.

A closer look

Find the work behind the skills.

Ask about engineering decisions, validation, experience, or project limitations. Each answer links back to its sources.

AI can make mistakes. Check the linked records. Your question is sent to Google Gemini.