QWED Protocol Verification

Version updated for https://github.com/QWED-AI/qwed-verification to version v5.1.0.

  • This action is used across all versions by 0 repositories.

Action Type

This is a Docker action.

Go to the GitHub Marketplace to find the latest changes.

Action Summary

QWED Protocol is a model-agnostic GitHub Action designed to verify the outputs, processes, and tool interactions of AI systems, including large language models (LLMs) and AI agents, before they are deployed to production. It prevents issues like hallucinations and inaccurate results by using multiple verification engines to ensure data trustworthiness and process determinism. This solution enhances reliability and security for AI-driven systems, enabling organizations to confidently use any LLM while mitigating risks associated with unverified outputs.

What’s Changed

QWED v5.1.0 extends the verification boundary beyond actions and into deterministic state governance.

This release introduces AgentStateGuard and ships a focused hardening wave that closes the post-v5.0.0 fail-open paths identified during adversarial review. The result is a stricter system that is more explicit about what it proves, harder to bypass under uncertainty, and safer under agentic workflows.

Why v5.1.0

This is a minor release because it adds a meaningful new capability while also tightening existing behavior across multiple trust boundaries.

Highlights:

  • new deterministic state governance via AgentStateGuard
  • fail-closed hardening across execution, tool approval, math verification, API semantics, and schema validation
  • clearer upgrade path for users who were still relying on legacy or ambiguous behaviors

Headline Feature: AgentStateGuard

AgentStateGuard expands QWED from action verification toward state and memory governance.

It adds:

  • strict structural validation for agent state payloads
  • semantic transition checks between current and proposed state
  • replay-resistant, monotonic transition enforcement
  • governed atomic commit behavior for approved state writes

This means agent state is no longer just “well-formed”; it can now be deterministically checked before becoming reality.

Additional Hardening

Execution and tool governance

  • CodeExecutor is now permanently hard-blocked as a legacy execution path
  • unknown tools are now denied by default instead of being auto-approved at low heuristic risk

Math and API semantics

  • verify_math() now enforces deterministic tolerance bounds
  • ambiguous /verify/math expressions now fail closed with BLOCKED
  • verify_identity() now rejects sampling-only agreement when no formal proof exists
  • verify_logic_rule() now fails explicitly instead of returning an ambiguous None

Schema enforcement

  • SchemaVerifier now fails closed when uniqueItems cannot be verified deterministically

Agent/runtime follow-through

  • progress-aware doom loop protection (LOOP-004)
  • continued infrastructure and config hardening
  • additional stats engine edge-case coverage
  • CodeQL and cleanup follow-up work

Upgrade Notes

Review these changes before upgrading:

  • CodeExecutor can no longer be used. Migrate direct imports to SecureCodeExecutor.
  • unknown tools now require explicit allowlisting
  • verify_math() may return BLOCKED for oversized or invalid tolerances
  • verify_logic_rule() now raises instead of returning None
  • verify_identity() sampling-only matches now return BLOCKED, not UNKNOWN
  • ambiguous /verify/math inputs now return BLOCKED with is_valid: false
  • uniqueItems verification failures now produce explicit schema errors instead of silent passes

Included Pull Requests Since v5.0.0

  • #124 feat(agent): add progress-aware doom loop guard (LOOP-004)
  • #126 security: harden configs, CI, and infrastructure — full audit fixes
  • #127 test(stats): add edge case coverage for statistics engine
  • #136 fix(codeql): resolve remaining syntax and test cleanup alerts
  • #137 Update contributors section in README
  • #139 feat: AgentStateGuard – full implementation (structural + semantic + atomic commit)
  • #149 fix: hard-block legacy CodeExecutor execution path
  • #150 fix: default deny unknown tool approvals
  • #151 fix: bound verify_math tolerance by computed magnitude
  • #152 fix: fail closed in verify_logic_rule
  • #153 fix: fail closed in verify_identity
  • #154 fix: fail closed for ambiguous math api inputs
  • #155 fix: fail closed on uniqueItems validation errors

Version Updates

  • qwed (PyPI): 5.0.0 → 5.1.0
  • qwed_sdk (Python): 5.0.0 → 5.1.0
  • @qwed-ai/sdk (NPM): 5.0.0 → 5.1.0

Full Changelog: v5.0.0...v5.1.0