Oleksandr Bezdieniezhnykh
c56d4584e6
Batch 87: 6 NFT-SEC blackbox scenarios + 5 helper evaluators + 75 unit tests + cumulative review batches 85-87. * AZ-436 NFT-SEC-01: cache-poisoning safety budget (AC-NEW-9); aggregate false_trust_count ≤ N×1e-6; zero-tolerance default. Canonical-only by default; E2E_NFT_SEC_01_RELEASE_GATE=1 unlocks full matrix. * AZ-437 NFT-SEC-02 + NFT-SEC-05: shared egress-observation evaluator (AC-NEW-10); SEC-02 = 0 packets to non-e2e-net over 5min replay; SEC-05 = DNS-blackhole sidecar healthy + lookup fails + UDP-53 silent. * AZ-438 NFT-SEC-03: AP-only signing rejection (AC-NEW-11); 3 sub-cases (unsigned/wrong-key/replayed) each reject ≤500ms + no position drift. * AZ-439 NFT-SEC-04: probe (always-run) = no-crash + deterministic decode outcome; ASan-fuzz (release-gate) = 0 findings ≥4h; AC-3 corpus floor informational only per spec. Verdict per-batch: PASS_WITH_WARNINGS (5 Low). Cumulative review for batches 85-87 (K=3 window) also PASS_WITH_WARNINGS with 5 cross-batch findings — recommends hygiene PBIs for write_csv_evidence duplication (13 helpers) and _resolve_fixture_path duplication (13 scenarios), plus new tickets for AZ-595 fixture builder + DNS-blackhole sidecar service. Also adds _docs/LESSONS.md documenting the Jira transition-ID lesson (always call getTransitionsForJiraIssue first, never memorize numeric IDs across sessions). Co-authored-by: Cursor <cursoragent@cursor.com>
1.1 KiB
LESSONS
Append-only ledger of lessons learned during the project. New entries go at the top. Each entry is one short bullet + a one-sentence "what changed".
2026-05-17 — Always call getTransitionsForJiraIssue before transitionJiraIssue
Trigger: In batch 87 (autodev step 10), I transitioned AZ-436..AZ-439 with transition.id="31" assuming = "In Progress" from stale memory. Read-back showed all four moved to Done instead (id 31 in this workflow = Done; In Progress = 21, In Testing = 32, To Do = 11). The mistake was caught by the tracker rule's mandatory read-back gate, fixed by re-transitioning to 21, and confirmed via second read-back.
What changed: Treat the transition ID as workflow-specific, not memorizable across sessions. Always query getTransitionsForJiraIssue first on the actual target issue (or one in the same project/workflow) and select the transition by name ("In Progress" / "In Testing" / "Done" / "To Do") — never by hard-coded numeric id. This is true even when you "remember" the IDs from a prior batch this same day, because the agent has no guarantee the workflow definition is stable.