docs(#1772): document dual test-result architecture and trust chain

[fullsend-ai-coder]

Add documentation to AGENTS.md covering the dual-path test validation architecture introduced in PR #1770:

• Architecture section: describes the parallel test validation paths
  (policy/release/test/ for pipeline task results vs
  policy/release/test_attestation/ for in-toto attestations) and their
  feature parity requirement.

• Architecture section: documents the trust verification chain in
  policy/lib/intoto/trust.rego and its fail-closed behavior.

• Common Change Patterns table: adds rows for test result validation
  changes and attestation trust verification modifications.

• Key Conventions: adds note about Sigstore/OCI mock patterns needed
  for test_attestation tests.

Co-Authored-By: Claude Opus 4.6 noreply@anthropic.com

---

Closes #1772

Post-script verification

• Branch is not main/master (agent/1772-doc-dual-test-arch)
• Secret scan passed (gitleaks — da5071a1d8a1e594609cd3c4c3ad6c640f85f6cf..HEAD)
• Pre-commit hooks passed (authoritative run on runner)
• Tests ran inside sandbox

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 8:13 PM UTC · Completed 8:21 PM UTC
Commit: 87c4a29 · View workflow run →

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

Flag	Coverage Δ
unit-tests	100.00% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[fullsend-ai-review]

Review — approve

PR: #1773 — docs(#1772): document dual test-result architecture and trust chain
Scope: 1 file changed (AGENTS.md), +21/−0

Summary

Clean documentation-only PR that adds architectural context about the dual test-result validation paths and trust verification chain to AGENTS.md. All changes trace directly to issue #1772 with no scope creep.

Dimensions reviewed

Dimension	Result
Correctness	✅ All 7 technical claims verified against source code
Security	✅ No secrets, injection patterns, or Unicode attacks
Intent & coherence	✅ Scope exactly matches issue #1772; no scope creep
Style & conventions	✅ Formatting, tone, and backtick usage consistent with existing AGENTS.md
Documentation currency	✅ All referenced file paths and identifiers verified as existing
Cross-repo contracts	⏭ Skipped — no exported interfaces modified

Technical accuracy verification

Every claim in the documentation was cross-checked against the actual source files:

• policy/release/test_attestation/ — exists, validates in-toto test-result attestations via intoto.verified_statements_by_predicate ✓
• policy/release/test/ — exists, validates pipeline task results consuming TEST_OUTPUT from SLSA provenance ✓
• policy/lib/intoto/trust.rego — exists, couples lib.intoto, lib.sigstore, lib.tekton, and OCI builtins. Header comment (lines 19–20) confirms the coupling ✓
• Fail-closed behavior — verified_statements (line 51) uses Rego comprehension semantics; failed conditions silently exclude statements. Tests (test_malformed_blob, test_unrecognized_statement_type) confirm this behavior ✓
• Test files — both test_attestation_test.rego and trust_test.rego exist and contain Sigstore/OCI mock infrastructure ✓
• Consumer deny rules — test_attestation.rego contains multiple deny rules operating on _test_attestations sourced from verified_statements ✓

No findings

No issues identified at any severity level. The documentation is technically accurate, well-placed within the existing document structure, and appropriately scoped to the linked issue.

---

Protected paths detected — this PR modifies files under one or more
protected paths. The review agent cannot approve PRs that touch these paths.
A human reviewer must approve this PR.

Protected files in this PR:

• AGENTS.md

---

Labels: PR adds documentation to AGENTS.md