ci(github-action)!: Update action misospace/pr-reviewer-action ( v2.2.1 ➔ v3.1.0 ) by mortyops[bot] · Pull Request #1905 · Aviator-Coding/home-ops

mortyops · 2026-10-01T00:03:01Z

This PR contains the following updates:

Package	Type	Update	Change
misospace/pr-reviewer-action	action	major	`v2.2.1` → `v3.1.0`

Release Notes

misospace/pr-reviewer-action (misospace/pr-reviewer-action)

`v3.1.0`

Compare Source

Features

tools: raise the native-loop budget and report a real cache hit ratio (#911) (bf937d0)

`v3.0.1`

Compare Source

Bug Fixes

release: track a source-vX.Y.Z anchor so release-please finds the last release (#907) (443bf19)
v3: read system-prompt-file from the base ref, not the run artifact dir (#905) (de5c628)

`v3.0.0`

Compare Source

⚠ BREAKING CHANGES

review: remove incremental-only escalation and review metadata state (#618) (#645)
review: remove carried findings and cross-run incremental evidence state (#617) (#644)
review: review_scope, escalate_on_dirty_baseline, effective_review_scope, previous_head_sha, and baseline_clean are removed; there is no replacement.

Features

ci: privilege-separated AI review for fork PRs (#748) (76d9baf)
context: anchor edits to their enclosing function (#790) (14a7cff)
context: consumers of changed keys and referenced counterparts (#795) (6376583)
context: current PR metadata outranks earlier discussion (#812) (f52f196)
deep-review: classifier-driven auto role selection (#633) (#655) (3a1ea6f)
deps: update dependency @types/node (24.13.6 → 24.19.0) (#760) (77203aa)
deps: update dependency esbuild (0.25.5 → 0.28.2) (#693) (3f6bfbf)
deps: update dependency typescript (5.8.3 → 7.0.2) (#696) (1327d75)
eval: blind adjudication tool for real-PR recall (#841) (#864) (7dd6fcb)
eval: counterexample-falsification scoring and fixtures (#757) (#758) (aaeac1f)
evals: A/B specialist execution shapes (#635) (#691) (ae8bb25)
evals: add historical semantic regression corpus (#627) (#646) (f86bf2c)
evals: add PR #654 execution-boundary fixtures (#659) (#664) (894f092)
evals: harvest maintainer findings into the human-findings corpus (#798) (#800) (4ad14e8)
evals: real-PR recall corpus (#782) (ea7649d)
platform: resolve tangled platform and normalize Spindle runtime context (#793) (b8b1aa8)
publish: resolve review threads of superseded managed reviews (#769) (6f5813a)
requirements: harness-authored verification obligations in the ledger (#796) (#818) (2a0d088)
review: add failure-path contract auditing to the correctness pass (#648) (ad79a2b)
review: budget and shape requests per tier (#668) (08a05b3)
review: gate unknown requirements (#647) (49190cb)
review: give smart escalation native tools (#665) (9334f02)
review: independent specialist corpus and output budget (#632) (#652) (082cf92)
review: make must_check a typed review obligation (#750) (#755) (5900cfc)
review: overlap specialist passes with CI gating (#634) (#654) (38b79bb)
review: state outstanding human change requests in every review (#774) (73246dd)
review: unresolved review threads as review context with dispositions (#770) (1591146)
routing: make post-primary smart escalation reviewer-requested only (#721) (#743) (9c13356)
tools: per-tier request budget inputs (#768) (3da29c5)
tools: stop rationing tool-loop evidence gathering (#794) (#797) (373529e)
tools: structured budget telemetry for the native loop (#709) (7c26e6d)
tools: tier-aware, exhaustion-aware native-loop request budgets (#707) (e2ca972)
transport: bound v3 model response buffering (#745) (#754) (55b5762)
v3: CI and specialists gate workloads (#807) (c504e51)
v3: clamp max_tokens to the provider's stated cap and retry once (#825) (7ba3463)
v3: consistent re-reviews: CI-aware skip and resolved-blocker gate (#812) (#817) (6fcbe1e)
v3: define kebab-case action contract (#687) (95c4039)
v3: deterministic context producers (#805) (eea90ba)
v3: end-to-end TypeScript orchestrator and v2 shadow comparison (#809) (#814) (d48b4f2)
v3: evidence providers, SARIF, image transport (#806) (4777297)
v3: make harness obligations opt-in (#796) (#831) (b7a6db4)
v3: migrate classification, role selection, and requirement ledger to TypeScript (#675) (#708) (39aa503)
v3: migrate enforcement, managed metadata, publishing, and outputs (#680) (#778) (db45c7e)
v3: migrate platform adapters and precheck path to TypeScript (#674) (#705) (567977e)
v3: migrate remaining #675 context scope to TypeScript (#716) (ff47ef3)
v3: opt-in claim falsification pass (#785) (#786) (e95fbe6)
v3: platform read seams (#803) (1e9dd0c)
v3: port change_anchors (#802) (7236aa4)
v3: port corpus assembly to TypeScript (#676) (#746) (37242e2)
v3: port routing, escalation, conversation, and native tool loop (#678) (#761) (7137b88)
v3: port the deep-review specialists logic (#776) (#780) (08bd1d1)
v3: port the model call layer to typescript (#703) (309aabf)
v3: production cutover to the TypeScript runtime (#706) (#815) (f94bef8)
v3: prompt and message layer (#804) (ebb231a)
v3: record tool-budget provenance in telemetry and the review marker (#847) (#865) (c794392)
v3: repository config, trim dead inputs (#784) (3cbfe6b)
v3: scaffold typed runtime (#689) (77736bf)
v3: size-scaled tool budget and honest partial coverage (#810) (#813) (eec5bc7)
v3: SSRF-safe linked-source fetch and render (#808) (5ca8d80)
v3: strict verdict derived from open findings and coverage (#811) (#816) (bf9ec54)
v3: tag findings outside the diff (#781) (9c9b103)
v3: thread re-emission keeps original severity; current metadata outranks earlier discussion (#812) (d21dc7d)
v3: typed gates, evidence, and lifecycle (#717) (709bb6f)
verdict: coverage, docs and style findings cannot block on their own (#772) (66cef8a)
verdict: decisive finding first (#773) (6ac94c5)

Bug Fixes

ci: wall-bound GitHub gh api polling attempts (#688) (d6889b4)
classifier: require a real untrusted-path surface for path handling (#756) (a3abef6)
classifier: stop misclassifying module imports as path handling (#720) (9931ac6)
dogfood: exercise all specialist roles (#667) (702675e)
dogfood: make the v3 shadow actually run (#819) (91df3a0)
enforcement: run thread settlement and carry original severity (#812) (7a7a35b)
eval: fail closed on an unusable pinned file manifest (#833) (ed873b2)
eval: fixtures are same-repo PRs; configurable review timeout (#771) (645feb6)
eval: key adjudication scenarios by corpus entry id (#893) (#894) (c5c1024)
eval: real-PR runs-per-mode, longer review timeout, bad corpus pin (#860) (062c498)
eval: seed pinned replays' file list from the pinned diff (6cc1113)
eval: seed pinned replays' file list from the pinned diff (#833) (c3db3c0)
evals: fail eval-harness when every run errors (#714) (05aaafe)
evals: harvester review follow-ups (#801) (88b94fc)
evals: retain diffs for specialists fixtures (#713) (fd4d8cd)
evals: triage the 13 anchor-check flags from #861 (#877) (60dbe1a)
eval: weekly summary reads per-mode rates from mode_summary (#715) (#744) (4b17206)
image: validate registry repository paths (#724) (864ffc1)
keep reviewed checkout off python sys.path (#788) (4c630a5)
model: guarantee at least one call attempt per tier (#887) (55eeaf8)
native-loop: consume recoverable verdict retry, no duplicate synthesis (#637) (#651) (345351c)
platform: body-edit cutoff via GraphQL lastEditedAt; reapply #812 rule post-CI (7a19eaf)
platform: one atomic body snapshot per metadata pass (#812 review) (ac09f43)
platform: valid PR-thread comment order (#631) (#649) (d45a273)
precheck: link issues referenced by PR title and non-closing body forms (#879) (e481bd6), closes #872
precheck: pin config reads to one fd (#723) (2ab2a3a)
publish: keep the endpoint URL out of published reviews (#836) (7a788da)
redact: use a non-alphanumeric marker for known-secret masking (#888) (939e0fc)
repo-map: avoid auth filename backtracking (#722) (d70bc95)
restore the reviewer's evidence gathering (v2 + v3) (#759) (339d998)
reviewer: keep sanitizer redaction from garbling committed source (#880) (791c85f)
run: streamed tool calls, tool cwd and specialist artifact root (#820) (515943e)
standards: resolve the standards file from the base ref, not PR head (#890) (59d93f2)
tests: clear .test-build and make git operations hermetic (#858) (27d4b95)
tool-loop: honour an empty ai_temperature on native_loop planning turns (#747) (8f3c0a6)
tooling: refuse redirects on token-bearing HTTP calls (#859) (860b56d)
tools: keep tool-harness.json valid under redaction and fail closed when unreadable (#899) (#900) (97f0b0d)
tools: scale loop round cap with the tool budget (#895) (#897) (a98bdf5)
transport: surface HTTP status and body detail on model-call errors (#862) (ec25b3f)
v3: mask secrets in model errors carried by a 200 reply (#868) (#869) (ae72527)
v3: normalize labeled-event shape so ai-review forces a re-review (#896) (241dd2d)
v3: publish the model-failure notice as request_changes (#863) (#866) (fb49242)
v3: run never reads artifacts from the PR checkout (#838) (#870) (6080648)
verdict: make non-blocking finding categories opt-in (#775) (1b83be8)
verdict: repair invalid escapes before accepting a partial candidate (#767) (d9b1a12)

Chores

clear code-scanning alerts before v3.0.0 (#891) (6c35674)
remove orphaned .candidates.json from repository root (#822) (#834) (20fd562)
v3: wave 2 teardown: freeze parity on v2 goldens, delete the remaining v2 code (#681) (#853) (a8c4d14)

Documentation

add a v2 to v3 upgrade guide (#857) (7a018e3)
agents: slim AGENTS.md to durable rules, move detail to docs (#751) (#753) (85ae3c3)
architecture: keep three_call specialist shape on benchmark evidence (#704) (#710) (4403d62)
runtime: choose composite Node boundary (#682) (3a362a2)
v3 examples and kebab-case input names (#848, #850, #851) (#849) (cd0d29c)
v3-migration: record the thread-severity and superseded-discussion change (04fa6ee)
v3: plan the v2 teardown file-by-file (#706) (#827) (61314cc)

Refactors

action: dedupe step env blocks into a shared $GITHUB_ENV export (#653) (00e7b01)
review: collapse the review corpus to one full-PR path (#643) (ec9e038), closes #616
review: remove carried findings and cross-run incremental evidence state (#617) (#644) (07fb9d4)
review: remove incremental-only escalation and review metadata state (#618) (#645) (a9dde07)
review: remove the review-scope selection seam (#615) (#638) (e2d74c2)
review: split render_linked_sources into single-purpose helpers (#640) (#650) (2af34b8)

`v3`

Compare Source

Bug Fixes

release: track a source-vX.Y.Z anchor so release-please finds the last release (#907) (443bf19)
v3: read system-prompt-file from the base ref, not the run artifact dir (#905) (de5c628)

`v2.5.0`

Compare Source

Features

evals: specialist-mode eval fixtures and effectiveness telemetry (#610) (#630) (0b1ce16)
review: add bounded specialist advisory-contract module (#612) (535f1b3)
review: add deep-review specialist advisory passes (#608) (#623) (8b4f314)
review: add requirement ledger with coverage credit (#624) (#628) (f2e5c3b)
review: feed specialist leads into final review synthesis (#609) (#629) (cad7827)
tangled: add ATProto CI session and record client (#587) (#611) (d3b6fa2)

`v2.4.0`

Compare Source

Features

change-anchors: extract deterministic change anchors from PR diffs (#582) (fc5f0d4)
context: add bounded PR thread context (#606) (e8cfa01)
context: add related code scanner (#600) (4ce23b8)
context: add repository map context (#599) (219c6fb)
context: wire related code into reviews (#601) (f635ee5)
evidence: normalize SARIF findings (#602) (51e31af)
evidence: wire local SARIF files into the existing evidence pipeline (#605) (a0651d2)
git_grep: add path scoping and explicit result limits (#596) (baa075b)
publish: add upstream_link_mode toggle for togithub.com links (#562) (5e6214a), closes #561
repo-map: add deterministic bounded repository-map builder (5764321)
tooling: add bounded find_files tool for filename/path discovery (#593) (73692b1)
tools: add bounded list_tree tool for repository discovery (#566) (9ce0066)
tools: add repo contents reader (#604) (5de3d65)

Bug Fixes

precheck: fail closed when the @ai-reviewer dismiss permission check can't run (#597) (15007ab)
repo-map: address PR review on three artifact-contract issues (f7cae56), closes #569

Chores

dogfood: raise native-loop budget to 4 rounds / 8 requests / 600s (#594) (56546a1), closes #565

`v2.3.3`

Compare Source

Bug Fixes

action: read fail_on_request_changes gate verdict from the output context (#558) (6cfa5bf), closes #557
corpus: gate the Tool Harness Findings section on harness output (#560) (243336a)

`v2.3.2`

Compare Source

Bug Fixes

config: let a single-provider configuration start (#555) (47385d4)

`v2.3.1`

Compare Source

Bug Fixes

carry-forward: propagate needs_full_review so unverifiable carried findings trigger a full review next run (#551) (845427e), closes #544
forgejo: bound gh CLI subprocess calls with a timeout (#548) (d7f5670), closes #537
forgejo: distinguish "unknown" permission from transport failure in preflight (#545) (dd27aaf), closes #539
forgejo: make the Authorized Integration JWT cache thread-safe (#547) (57944ac), closes #538
precheck: re-review on an unchanged diff when the blocker was CI state (#550) (267b862)
precheck: wire the @ai-reviewer dismiss directive into production (#554) (dd13151)

Chores

tools: rename dead tool_planning_* inputs to their actual purpose (#552) (75eefbc), closes #540

Refactors

publish: extract Publish step inline bash dispatcher to scripts/publish.sh (#553) (5b6b2cf), closes #541

`v2.3.0`

Compare Source

Features

carry_forward: let a maintainer dismiss a carried finding (#534) (#535) (623ec46)

Bug Fixes

classifier: stop matching bare route.<ext> in public_route_changes (#532) (e106089), closes #531
tools: preserve standards evidence (#542) (d2412dc)

Configuration

📅 Schedule: (in timezone America/New_York)

Branch creation
- At any time (no schedule defined)
Automerge
- At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.

If you want to rebase/retry this PR, check this box

This PR has been generated by Mend Renovate CLI.

….1 ➔ v3.1.0 ) | datasource | package | from | to | | ----------- | ---------------------------- | ------ | ------ | | github-tags | misospace/pr-reviewer-action | v2.2.1 | v3.1.0 |

mortyops Bot added renovate/github-release renovate/github-action type/major area/github labels Oct 1, 2026

ci(github-action)!: Update action misospace/pr-reviewer-action ( v2.2…

a71b056

….1 ➔ v3.1.0 ) | datasource | package | from | to | | ----------- | ---------------------------- | ------ | ------ | | github-tags | misospace/pr-reviewer-action | v2.2.1 | v3.1.0 |

mortyops Bot force-pushed the renovate/misospace-pr-reviewer-action-3.x branch from b37932b to a71b056 Compare October 1, 2026 04:03

mortyops Bot changed the title ~~ci(github-action)!: Update action misospace/pr-reviewer-action ( v2.2.1 ➔ v3.0.1 )~~ ci(github-action)!: Update action misospace/pr-reviewer-action ( v2.2.1 ➔ v3.1.0 ) Oct 1, 2026

Navigation Menu

ci(github-action)!: Update action misospace/pr-reviewer-action ( v2.2.1 ➔ v3.1.0 ) - #1905

ci(github-action)!: Update action misospace/pr-reviewer-action ( v2.2.1 ➔ v3.1.0 )#1905
mortyops[bot] wants to merge 1 commit into
mainfrom
renovate/misospace-pr-reviewer-action-3.x

mortyops Bot commented Oct 1, 2026 •

edited

Loading

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

0 participants

Conversation

mortyops Bot commented Oct 1, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Release Notes

Features

Bug Fixes

⚠ BREAKING CHANGES

Features

Bug Fixes

Chores

Documentation

Refactors

Bug Fixes

Features

Features

Bug Fixes

Chores

Bug Fixes

Bug Fixes

Bug Fixes

Chores

Refactors

Features

Bug Fixes

Configuration

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

0 participants

mortyops Bot commented Oct 1, 2026 •

edited

Loading