📊 Triage digest — week of 2026-07-13

[claude]

📊 Triage digest — week of 2026-07-13

Issues: 15 opened (5 Raven sd-XXX findings: sd-003, sd-005, sd-007, sd-008, sd-009) · 3 closed (1 refuted-Raven auto-close, 1 exact-duplicate auto-close, 1 completed via merged fix PR)
PRs: 12 opened · 4 merged · 0 closed unmerged · 8 still open from this week's cohort (14 open overall)
REVIEW_REQUIRED: 0 (review:required search returns none — this repo has no branch-protection required-reviewers configured) · no prior digest exists to compare against, so this is the baseline week

• The Raven lane processed 5 findings this week: sd-003 was refuted and auto-closed (sd-003: The Algolia docs index excludes the auto-generated RPC-method and #13 — the indexed Structure→Pagination page already lists the getTransactions/getLedgers limits the finding claimed were missing); sd-008 was confirmed P1 (sd-008: The Software Versions page still presents Protocol 27 as Testnet-only after Mainnet activation #11 — Protocol 27 still shown Testnet-only) and already has an open fix (Add Protocol 27 Mainnet section to software versions #22, currently needs-changes); sd-005 (sd-005: The docs' agentic-payments coverage documents x402 and MPP mechanics in depth #14) and sd-009 (sd-009: The docs omit SEP-6's canonical interactive-component deprecation status #10) were confirmed P2; sd-007 (sd-007: The state-archival docs now correctly describe Protocol 23 automatic #12) hasn't been triaged at all yet — no labels, no comments.
• One duplicate slipped in from outside the Raven lane: software-versions.mdx still shows Protocol 27 as Testnet-only after mainnet launch #24 reported the same Protocol-27-Testnet-only defect as sd-008: The Software Versions page still presents Protocol 27 as Testnet-only after Mainnet activation #11/sd-008 and was closed as an exact duplicate. Worth a policy note — the timeline on software-versions.mdx still shows Protocol 27 as Testnet-only after mainnet launch #24 shows the agent (claude[bot]) applying triage:approve-close to its own finding and the executor closing it 10 seconds later, rather than a maintainer approving a standing triage:close-candidate proposal as the two-step close policy specifies.
• Three typo/import auto-fixes went end-to-end this week: agent labels triage:autofix-candidate → opens a fix PR → PR gets auto-merge-candidate → merges. That closed Typo: "recieves" should be "receives" in the SEP-6 wallet doc #30 (via Fix typo: 'recieves' -> 'receives' in SEP-6 wallet doc #32) and separately landed Fix broken py-stellar-base imports in Hardhat migration example #23 and Fix typo: recieve → receive in SEP-31 getting started #28 the same way.
• Typo: "occured" should be "occurred" in transactions result-codes doc #29 (typo, triage:autofix-candidate) is stuck: re-verified three separate times with no state change, a maintainer asked @copilot can you create a branch? on 2026-07-14, and three days later there's still no PR.
• No human overrides of agent verdicts this week. The only close/reopen pair in the window (Smoke test: getting-started guide references stellar-cli v22 output #8) was the maintainer's own smoke test — both actions by the same human, seconds apart, before the agent ever touched the issue.

Needs human

• Open P1 (ex-P0), no merged fix yet:
  • #17 — Fix broken JS SDK code snippets (draft fix in progress: #31)
  • #11 — sd-008: Protocol 27 shown Testnet-only (fix open in #22, currently needs-changes)
  • #9 — Stellar RPC OpenRPC spec reports stale info.version
• triage:close-candidate idle > 7 days: none — only #8 currently carries the label, and it's 3 days old.
• Other stalled item worth a look: #29 — autofix-candidate typo, re-verified 3x, no PR after a maintainer's @copilot request 3 days ago.