Skip to content

Releases: sunnnn2005/data-quality-agent

Release list

Data Quality Agent v0.3.0

Data Quality Agent v0.3.0 Latest
Latest

Choose a tag to compare

View all tags
@sunnnn2005 sunnnn2005 released this 13 Aug 12:05
v0.3.0
b9ee041

Data Quality Agent v0.3.0

Data Quality Agent v0.3.0 packages the project as a stronger AI-engineering resume signal. The release focuses on proving that the system is an LLM agent with controlled tools, state, evidence-backed reporting, safety boundaries, and public verification artifacts. It still avoids claims that are not proven yet: no external users, customer feedback, enterprise production deployment, or paid-model benchmark results are claimed.

What changed

  • Added a CI-verified LLM agent capability matrix with 13 implemented checklist items, 4 partial maturity areas, 7 allowed tools, and explicit not-claimed production adoption.
  • Expanded the live project scorecard to connect the public demo, release, container image, 88 passing tests, 47 verified resume claims, 11 reviewer paths, and application evidence links.
  • Updated recruiter-facing and application-facing artifacts so project claims are backed by public URLs instead of unsupported resume language.
  • Kept the project honest about traction: GitHub stars, forks, feedback issues, confirmed users, and reproducible feedback remain tracked from public sources instead of estimated.
  • Preserved the read-only PostgreSQL agent route, deterministic fallback behavior, agent safety boundaries, trace observability, business-rule retrieval, and report guardrails from earlier releases.

Verified evidence

Verified status

  • 88 automated tests pass locally and in GitHub Actions.
  • 47 resume-ready claims are listed in the outcome evidence manifest.
  • 13 LLM-agent checklist capabilities are implemented in the public matrix.
  • 4 maturity areas are intentionally marked partial.
  • 1 area is explicitly not claimed: external users, customer feedback, enterprise production deployment, paid model benchmark results, and autonomous write actions.
  • Public metrics currently track 0 stars, 1 fork, 0 confirmed external users, and 0 external feedback items.

Boundaries

  • No external users are claimed yet.
  • No customer feedback is claimed yet.
  • No enterprise production deployment is claimed yet.
  • The PostgreSQL demo uses seeded local support-ticket data and a read-only adapter.
  • External model calls remain opt-in through OPENAI_API_KEY.
  • The project is resume-ready as public engineering evidence, not as a production customer adoption claim.
Assets 2
Loading

Data Quality Agent v0.2.0

Data Quality Agent v0.2.0

Choose a tag to compare

View all tags
@sunnnn2005 sunnnn2005 released this 13 Aug 05:03
v0.2.0
e578055

Data Quality Agent v0.2.0

Data Quality Agent v0.2.0 turns the first public demo into a stronger resume-evidence release. The core product is still intentionally conservative: it demonstrates a realistic business data-quality workflow without claiming external users, customer feedback, or enterprise production adoption.

What changed

  • Added a PostgreSQL-backed LLM agent route: POST /postgres/support-tickets/agent-report.
  • Reused the read-only support-ticket database adapter for both deterministic reports and agent reports.
  • Added safe disabled fallback when the LLM agent route is called without OPENAI_API_KEY.
  • Expanded the public outcome evidence manifest with a verified postgres-agent-route claim.
  • Expanded scheduled public evidence health checks from 6 to 7 checks, including source verification for the PostgreSQL agent route.
  • Updated adoption metrics and history to reflect 52 passing CI tests.
  • Kept feedback and user metrics at an honest zero baseline until public evidence exists.

Verified evidence

Verified status

  • 52 automated tests pass in GitHub Actions.
  • 7 public evidence health checks pass.
  • Public metrics currently track 0 stars, 1 fork, 0 confirmed external users, and 0 external feedback items.
  • The support-ticket case study still verifies duplicate ticket IDs, missing routing fields, negative amounts, and amount outliers.

Boundaries

  • No external users are claimed yet.
  • No customer feedback is claimed yet.
  • No enterprise production deployment is claimed yet.
  • The PostgreSQL demo uses a seeded local support-ticket table and a read-only adapter.
  • External model calls remain opt-in through OPENAI_API_KEY.
Loading

Data Quality Agent v0.1.0

Data Quality Agent v0.1.0

Choose a tag to compare

View all tags
@sunnnn2005 sunnnn2005 released this 13 Aug 03:31
v0.1.0
af84cd9

Data Quality Agent v0.1.0

Data Quality Agent v0.1.0 is the first public release of the project.

What is included

  • Public demo: https://sunnnn2005.github.io/data-quality-agent/
  • Support-ticket case study: docs/support-ticket-case-study.md
  • Machine-verified artifact: docs/verified-support-ticket-result.json
  • Read-only PostgreSQL Docker Compose demo
  • OpenAI-compatible LLM tool-calling agent
  • Source-cited business-rule retrieval
  • Bounded CSV business-data upload
  • Auditable run traces
  • Three-scenario evaluation harness
  • Public feedback issue template

Verified demo result

The support-ticket demo is verified by scripts/verify_support_ticket_demo.py in CI:

  • status: FAIL
  • quality score: 24
  • row count: 8
  • finding count: 5
  • expected checks:
    • duplicate_primary_key
    • missing_values
    • negative_amount
    • numeric_outliers
  • expected business rules:
    • support_tickets:R1
    • support_tickets:R2
    • support_tickets:R3
    • support_tickets:R4

Test status

  • 39 automated tests pass locally and in GitHub Actions.
  • CI runs both the deterministic test suite and the support-ticket demo verifier.

Boundaries

  • The default path does not require paid APIs or model keys.
  • External model calls are opt-in through OPENAI_API_KEY.
  • PostgreSQL access is disabled by default and uses a read-only adapter.
Loading