Resources
Curated Runs
Vendor-neutral evidence packs for third-party verification. Packs are evaluated against their declared ruleset and can be independently recomputed locally.
Runs are curated references from runsets.yaml. Only runs with an adjudication bundle have deterministic verdict_hash and can be rechecked.
Cross-Vendor Evidence Spine
- •Scenario Focus: LG-01 (Single Agent Lifecycle)
- •Ruleset: ruleset-1.0 (presence-level validation)
- •Important: LG-02~05 PASS results indicate artifact presence only, not semantic correctness for those flows.
Scope NoteNOT_ADJUDICATED indicates the run has no applicable ruleset reference or lacks required artifacts for its declared ruleset. v0.4 packs use ruleset-1.2 with 12 semantic invariant clauses.
Semantic Invariant Packs (v0.4)
ruleset-1.212 clauses · 4 domains| Run ID | Substrate | Status | Verdict Hash | Actions |
|---|---|---|---|---|
| arb-d1-budget-fail-deny-missing-gate-v0.4 | fixture | NOT_ADJUDICATED | 32c0d954a9bc3cb1... | |
| arb-d1-budget-fail-outcome-invalid-v0.4 | fixture | NOT_ADJUDICATED | de4cfcefa2dab530... | |
| arb-d2-lifecycle-fail-post-terminal-exec-v0.4 | fixture | NOT_ADJUDICATED | f6ae7754f881ea6c... | |
| arb-d2-lifecycle-fail-terminal-state-invalid-v0.4 | fixture | NOT_ADJUDICATED | 5d35c02384510395... | |
| arb-d3-authz-fail-deny-no-confirm-v0.4 | fixture | NOT_ADJUDICATED | c2bd785a904696c8... | |
| arb-d3-authz-fail-sra-incomplete-v0.4 | fixture | NOT_ADJUDICATED | a5daf338965bd12e... | |
| arb-d4-termination-fail-reason-invalid-v0.4 | fixture | NOT_ADJUDICATED | 959f6683ba9dcb1a... | |
| arb-d4-termination-fail-uncontrolled-recovery-v0.4 | fixture | NOT_ADJUDICATED | ad7fc6c84f70927c... |
Note: Curated Runs = indexed & contract-governed. Legacy runs (Autogen, Magentic One, etc.) in releases/ are archived and not part of v0.5 freeze.
Four-Domain Packs (v0.3)
ruleset-1.1D1/D2/D3/D4 domains| Run ID | Substrate | Status | Verdict Hash | Actions |
|---|---|---|---|---|
| arb-d1-budget-fail-fixture-v0.3 | fixture | NOT_ADJUDICATED | 8e7e5f62a9698225... | |
| arb-d1-budget-pass-fixture-v0.3 | fixture | NOT_ADJUDICATED | 792256da104fbb59... | |
| arb-d2-lifecycle-state-fail-fixture-v0.3 | fixture | NOT_ADJUDICATED | 58d47ba99fce73ee... | |
| arb-d2-lifecycle-state-pass-fixture-v0.3 | fixture | NOT_ADJUDICATED | c01b5c9acf3c3696... | |
| arb-d3-authz-decision-fail-fixture-v0.3 | fixture | NOT_ADJUDICATED | 664e9c64e984ac47... | |
| arb-d3-authz-decision-pass-fixture-v0.3 | fixture | NOT_ADJUDICATED | 62911375fae4d5fd... | |
| arb-d4-termination-recovery-fail-fixture-v0.3 | fixture | NOT_ADJUDICATED | d9cc0d51523d3178... | |
| arb-d4-termination-recovery-pass-fixture-v0.3 | fixture | NOT_ADJUDICATED | d28cec06ea565d45... |
Note: Curated Runs = indexed & contract-governed. Legacy runs (Autogen, Magentic One, etc.) in releases/ are archived and not part of v0.5 freeze.
GoldenFlow Packs (v0.2)
ruleset-1.0npm run vlab:recheck-hash| Run ID | Substrate | Status | Verdict Hash | Actions |
|---|---|---|---|---|
| admission-not-admissible-01 | fixture | NOT_ADMISSIBLE | bb50519dec88035d... | |
| gf-01-a2a-fail | a2a | NOT_ADJUDICATED | 94d802b0923db249... | |
| gf-01-a2a-official-v0.2 | a2a | NOT_ADJUDICATED | 0000000000000000... | |
| gf-01-a2a-pass | a2a | NOT_ADJUDICATED | 9e3f137c7ff250fe... | |
| gf-01-adjudicated-fail-01 | fixture | ADJUDICATED | 79e5b0713db3f2b7... | |
| gf-01-langchain-official-v0.2 | langchain | NOT_ADJUDICATED | 0000000000000000... | |
| gf-01-langchain-pass | langchain | ADJUDICATED | 4372ae42ce357fdb... | |
| gf-01-mcp-official-v0.2 | mcp | NOT_ADJUDICATED | e7f5b31e40d6a645... | |
| gf-01-mcp-pass | mcp | ADJUDICATED | 3db969b9cdda1f0a... | |
| gf-01-smoke | fixture | ADJUDICATED | 6afde43633358e9a... | |
| gf-02-fail | fixture | NOT_ADMISSIBLE | 0dbf54c129f68274... |
Note: Curated Runs = indexed & contract-governed. Legacy runs (Autogen, Magentic One, etc.) in releases/ are archived and not part of v0.5 freeze.