↓ Skip to main content
tomrochette.com
  1. Agents/
  2. Agents log/

Agents log 2026-11-14

Author
Tom Rochette
History / Edit / PDF / EPUB / MD
Created: / Updated: / Status: finished / 36 minutes read (~7127 words)
agents 122
Table of Contents

(daily refresh, 1)
#

  • goose: rotation re-verification, references all live; appended the v1.58.0, v1.58.1, and v1.59.0 release bullets the Status prose carried without them, added v1.59.1 (November 14, notes pending), push date 2026-11-14.
  • juggler: rotation re-verification, references all live; appended the v0.8.0 release bullet (November 8) the prose carried.
  • nanocoder: rotation re-verification; repointed the dead scheduler.md reference to docs/features/skills.md, npm trailing month 10,391 (October 15 to November 14), v1.32.0 still newest, push 2026-11-14.
  • onecli: rotation re-verification, references and pricing tiers live and unchanged; appended the v2.20.0 release bullet (November 9) the prose carried.
  • warp-agent-cli: rotation re-verification, references and pricing page live and unchanged; CLI docs last-updated moved to November 11, 2026.
  • zcode: rotation re-verification, references all live; v3.17.2 still the newest desktop release, push 2026-11-13.
  • claude-code: recorded the train reaching v2.1.325 (November 13): Code Mode opt-in and plugin version-range gating in v2.1.323, zstd security fix and subagent context-sizing fixes in v2.1.325; repaired the Changes section’s ordering.
  • codex: recorded rust-v0.169.0 as the latest stable (November 13).
  • copilot-cli: recorded the stable dist-tag reaching 1.0.112 (November 13) and refreshed the trailing-month downloads to 6,944,357.
  • gemini-cli: appended the v0.68.0 tagged-release bullet (November 10) the Status prose carried.
  • junie: recorded the installer train continuing through 4131.1 (November 13).
  • mimo-code: recorded the v0.1.17 release (November 13) and refreshed the push date to 2026-11-13.
  • om-o: recorded the train reaching v5.1.70 (November 14) and appended the v5.1.65 to v5.1.69 bullets, push date 2026-11-14.
  • opencode: recorded the v2 train reaching v2.3.0 (November 13) and the v2.2.7 bullet, default branch pushed 2026-11-14.
  • longhorizon-harness: corrected the commit-silence claim: default-branch commits on October 21, 23, 26, and a merged MCP-server-integration PR on November 11, all unreleased; caution now reads revived-but-unshipped.
  • command-code: recorded the npm train reaching 1.89.1 (November 13) and refreshed the monthly downloads to 196,115.
  • kimi-code: K3 OpenRouter reading re-checked unchanged at $0.192 in / $16.00 out (cache reads $0.182) on 2026-11-14.
  • ag2: rotation re-verified, release line and PyPI tracking stand at 1.3.1, push refreshed to 2026-11-14.
  • agent-teams: rotation re-verified, v2.18.7 stands, push refreshed to 2026-11-14.
  • autogen: rotation re-verified, archive declaration stands, successor MAF push refreshed to 2026-11-14.
  • autogpt: rotation re-verified, v0.11.2 stands, hosted pricing re-fetched unchanged, push refreshed.
  • bernstein: rotation re-verified, v3.24.1 stands on GitHub and PyPI, push refreshed.
  • chorus: rotation re-verified, recorded v0.24.1 (2026-11-14, default-on run sync), npm counts refreshed to 55 versions and 5,591 October downloads.
  • claude-squad: rotation re-verified, the eighty-five-day quiet spell broke with a single 2026-11-13 fix commit, Homebrew refreshed to 5,110 installs.
  • clawteam: rotation re-verified, dormancy stands.
  • concord-mcp: rotation re-verified, v0.22.0 stands, npm trailing-month downloads 4,131.
  • conductor: rotation re-verified, wind-down stands, cloud read-only window arrived on schedule 2026-11-13.
  • crystal: rotation re-verified, death record stands.
  • dmux: rotation re-verified, ninety days quiet, npm 861 for the month ending 2026-11-13, watch-list stands.
  • frink: rotation re-verified, 0.4.0 stands, default branch resumed 2026-11-09.
  • mastra: recorded @mastra/core 1.83.0 (2026-11-13, resumable agent streams on every Agent, built-in MCP server).
  • agent-orchestrator: recorded stable v0.15.0 (2026-11-13).
  • agent-swarm: recorded v2.13.1 (2026-11-14) at the train head.
  • buzz: recorded desktop v0.5.30 (2026-11-14) after the five-week desktop gap.
  • langgraph: recorded 1.2.22 (2026-11-14).
  • multica: recorded v0.9.0 (2026-11-13).
  • omnara: recorded omnarad and cluster v0.2.23 (both 2026-11-13).
  • openharness: recorded CLI line through v0.6.56_cli (2026-11-13).
  • openrig: recorded v0.7.14 (2026-11-13).
  • pandacode: recorded v0.3.125 (2026-11-14) adding Cursor CLI over the Agent Client Protocol.
  • ruflo: recorded v3.77.1 (2026-11-13, AgentDB embedding-model fix).
  • sim: recorded v0.9.57 (2026-11-14).
  • orchestration: pushed dates refreshed to the default-branch latest on agent-manager, cli-agent-orchestrator, commonly, gascity, happy-coder, nodeterm, orkas, parallel-code, paseo, t3code, yao.
  • orchestration: matrix Current-status row synced to twelve new release heads.
  • continue: rotation re-verification, all cited URLs re-fetched live, acquisition notice, read-only README (Apache-2.0, pushed November 13, not archived), Ollama guide, and both funeral threads stand, marketplace counter refreshed to 4,556,963 installs.
  • kiro: rotation re-verification recorded the November 10-12 changelog wave (Kiro Mobile for Android, IDE 1.4.178 and 1.4.154, CLI 3.3.0, 3.2.1, 3.2.0, Grok 4.7 experimental, Enterprise Management preview with Kiro Crew LTS, CloudTrail audit logging, credit pooling with Okta and Entra ID), tier prices unchanged.
  • visual-studio-2026: rotation re-verification recorded the November update 18.12.0 (new Copilot Chat default for eligible accounts, AGENTS.md at solution root plus organization guidance, per-skill controls, MCP management, AI Git conflict resolution, Change Groups), subscription ladder unchanged.
  • vibeyard: release train resumed November 13 after ten weeks with v0.3.9 (no written notes) and v0.3.10 (pipeline fix only), vibeyard-ctl now documented in the README, Windows installers added, status and cautions rewritten.
  • codeg: recorded v0.44.0 (November 13, uncommitted task changes joining PRs and MRs, Superpowers 7.0.2 and scientific-agent-skills 2.72.0 updates, bundled-agent bumps) as the new latest release.
  • vscode-copilot: recorded 1.145.3 (November 10) as the servicing patch the prior pass’s chain skipped, 1.146.0 stands current.
  • waku: rotation re-verification, all seven cited URLs re-fetched live, repository still dormant since October 2, author record refreshed.
  • surface-feature-matrix: corrected the Continue Open source cell to ✓ Apache-2.0 as an as-they-were snapshot (license re-verified against the GitHub API).
  • surface-feature-matrix: moved the Visual Studio 2026 AGENTS.md cell to solution-root AGENTS.md support on the 18.12 release, JetBrains stays the lone vendor-convention holdout.
  • surfaces: bb, cc-haha, CodeNomad, Nimbalyst, OpenChamber, and Zed checked on their pushed signals with no release or status movement to record.
  • surfaces: OpenBitFun’s Cowork-style general-workspace rejection holds on the ninth q#14 re-read (README re-fetched 2026-11-14, still four harness modes, phone remote control, Mini App marketplace, general-workspace pitch).
  • surfaces: no credible surfaces-dedicated source has appeared (GitHub search re-run, same low-star candidates).
  • assistant-runtimes: volatile status refresh across thirteen notes after the batched GitHub API pass, no release moves (every latest release matched its note)
  • assistant-runtimes: anything-llm - status refreshed, 338 open issues, pushed 2026-11-13
  • assistant-runtimes: astrbot - status refreshed, 1,217 open issues and PRs, pushed 2026-11-14
  • assistant-runtimes: eigent - status refreshed, 248 open issues, pushed 2026-11-13
  • assistant-runtimes: hermes - status refreshed, 53,388 open issues, pushed 2026-11-14
  • assistant-runtimes: langbot - status refreshed, 137 open issues, pushed 2026-11-14
  • assistant-runtimes: nanobot - status refreshed, 869 open issues, pushed 2026-11-14
  • assistant-runtimes: nanoclaw - status refreshed, 1,065 open issues as of 2026-11-14
  • assistant-runtimes: octop - status refreshed, 699 open issues, pushed 2026-11-14
  • assistant-runtimes: open-webui - status refreshed, 328 open issues, pushed 2026-11-14
  • assistant-runtimes: openclaw - status refreshed, 8,188 open issues
  • assistant-runtimes: openwork - status refreshed, 425 open issues and PRs, pushed 2026-11-14
  • assistant-runtimes: openworker - rotation re-verification, v0.3.1 still the latest desktop release and skills-v0.1.0-beta.2 still newest, OpenShell pin and approval model intact, Windows story corrected (official site now lists only a Mac download), tag normalized to assistant-runtimes
  • assistant-runtimes: picoclaw - no change, fresh counts matched the note
  • assistant-runtimes: private-gpt - rotation re-verification, v1.0.1 still latest, all claims unchanged, status moved to pushed 2026-11-13 with 12 open issues and PRs
  • assistant-runtimes: qwenpaw - status refreshed, 1,165 open issues and PRs
  • assistant-runtimes: zeroclaw - status refreshed, pushed 2026-11-13 with 1,217 open issues
  • assistant-runtimes: assistant-runtimes-feature-matrix - re-verification moved six status cells and matching prose counts
  • cross-category: the 2026-11-14 stars scan finds no starred_at past the 2026-11-08T06:11:27Z checkpoint (newest entry is the checkpoint itself), checkpoint rewritten unchanged
  • cross-category: q26/q114/q118/q121/q111/q116/q126 dedicated-list evidence refreshed 2026-11-14: autoresearch reads yibie/awesome-autoresearch 795 pushed 11-14 and handsome-rich/Awesome-Auto-Research-Tools 1,292 pushed 11-13 (both maintenance commits only, checked at commit level) against webfuse-com 2,591 (stale 09-21) and AI4Scientist/awesome-autoresearch 188 (pushed 10-28); deep-research largest DavidZWZ/Awesome-Deep-Research 875 unchanged stale since 07-23; andyrewlee/awesome-system-one 60 stars pushed 11-12 with a +19-entry recursive sweep and a Gemini Decisions API addition; awesome-jev head yibie/awesome-jev 2,585 pushed 11-14; AnotiaWang/awesome-decision-models 715 pushed 11-13 (Umpirix and OnlyRead additions); cuihuan/awesome-ai-gateway 156 and mahmoudsajjadi/awesome-ai-coding-tools 15 both pushed 11-13, both below the wired sources’ floor; kyrolabs/awesome-ade 27 pushed 11-11; all stay unwired or unmapped pending their owner calls

(daily refresh, 2)
#

  • a2a: re-verified, spec at v1.0.1, repository pushed 2026-11-13 on docs-only partner additions, official CLI train at v0.4.0 with 11-12 to 11-13 default-branch fixes and features.
  • a2ui: re-verified, tagged train unchanged (2026-11-10 Python and 2026-11-11 Dart waves), spec table v1.0 current re-confirmed, React joined the official SDK list, pushes 11-12 to 11-14 renderer and docs-site work.
  • acp: re-verified, no new tag (v1.17.0 with schema-v1.31.0 and schema-v2.0.0-alpha.15 stand), 11-13 to 11-14 unstable-v2 default-branch work, agents list re-counted at 51.
  • ag-ui: train moved to release/2026-11-13 (@ag-ui/langgraph 0.0.46, Mastra registerCopilotKit for CopilotKit 2.x), downloads refreshed to about 25.0M combined in the month ending 2026-11-14.
  • mcp: re-verified, spec revision 2026-07-28 stands (newest dated folder unchanged), docs and CI commits since 11-12 including EU Cyber Resilience Act stewardship guidance.
  • personal-agent-protocol: design-partner roster expanded 43 to 103 on 2026-11-13 (technology providers folded into companies, Manus added), draft 1.0 in progress with six breaking changes, repository pushed 2026-11-13.
  • uhp: train through v0.41.16 (2026-11-14), spec changelog gained the ETag/If-Match addition within the 2026-10-19 fifth version, suite stays at 2026.10.19.post6.
  • agents-md: rotation re-verification, every cited URL live, site still 60k+ carrying projects, repository pushed 2026-09-10, mod directory and changelog entries unchanged (newest 2.1.325).
  • anp: rotation re-verification, spec hub still ANP 1.2, footprint unchanged at 2/1/1 points, OSSA survey unchanged, default branch quiet since 2026-10-22, companion anp pushed 2026-11-13.
  • protocols: matrix stack-thesis count corrected thirteen to fourteen, AG-UI and Poppy adoption prose refreshed, all rows carry the header’s 15 cells, columns still title-sorted; no membership change.
  • protocols: internal link check clean across all notes, matrix, and the category index; no type fields; front matter parses.
  • context-engines: link, front-matter, section-order, and as-of checks pass across all 21 notes; matrix (21 columns) and category index (21 entries) match membership, sorted case-insensitively, cell-count parity holds, and the Kind row verified cell-by-cell against every member note; no membership change
  • context-engines: entroly - main 2026-11-13 batch recorded (signed offline-verifiable proof bundles, opt-in proxy header forwarding, pricing-catalog hardening); PyPI and npm unchanged at 1.0.88
  • context-engines: headroom - main 2026-11-14 security sweep recorded (PRs #4848-#4870: proxy credential hygiene, memory-tool scoping, Host/Origin checks, OAuth redirect refusal); no release since v0.44.0
  • context-engines: jevgrep - main 2026-11-12 CLI batch recorded (file-category filter flags, shebang TS runtime detection); npm trailing-month 27,373; v0.8.5 still latest
  • context-engines: repowise - release v0.61.0 (2026-11-13) shipped the recorded merge batches; PyPI at 0.61.0
  • context-engines: repomix - rotation re-verified, all cited URLs intact; 4,907 commits and 425,001 trailing-month npm recorded; commits since v1.19.0 dependency-only
  • context-engines: vanna - rotation re-verified; archive banner and the three pricing tiers unchanged; scmagazine citation stands on the bot-wall ruling
  • context-engines: augment-code - rotation re-verified; pricing plans, no-free-tier, and newest-post claims unchanged; intent-hq pushed 2026-11-14
  • context-engines: the remaining volatile signals resolved non-material under standing rulings - context7 (docs-only default-branch commits), graft and sol-pi (pushes off main), toon (VS Code chores), wrenai (single unreleased SDK fix), rtk (dependency bumps and small unreleased fixes, dev-1.2.1 train continuing); no notes touched
  • context-engines: category fit re-checked: all 21 members remain context engines, no pivots
  • agent-plugins: rotation re-verification held all v1.0.0 claims and the Status gained the v1.1.0 draft’s 11-13 portable-hooks merge (PR #89) and 11-14 name-validation and MCP-cwd fixes, published site still v1.0.0
  • agent-native: volatile refresh moved npm to 0.4.7 (published 2026-11-13) with pinned downloads 4,315 (11-08 to 11-14) and the pack re-counted at 25 skills
  • opencode-skills-and-plugins: rotation re-verification held all claims, the release train moved to v2.3.0 (2026-11-13), and the frontmatter clause was corrected to all-optional at runtime
  • skillspector: volatile refresh moved main to 2026-11-14 (five PRs, MiniMax M3.1 Flash defaults and nv_build default z-ai/glm-5.3-flash among them) and open issues to 145
  • snyk-agent-scan: volatile refresh moved the default branch to 2026-11-13 (retry hooks on transient errors, certificate-error fix)
  • skills: anthropic-agent-skills volatile check held (default branch still at the 2026-11-06 commit, the pushed_at move is side-branch activity on the 2026-11-01 ruling)
  • chonkie: pushed 2026-11-13; pypistats frozen 1,255,276 window refreshed to 1,500,669 (as of 2026-11-14); PyPI 1.7.0 still latest.
  • docling: PyPI 2.150.1 (2026-11-14, 241 releases); pushed 2026-11-14; trailing-month downloads re-read 4,037,393, the frozen 2,909,751 window ended, pepy badge corroborates 4M.
  • knowhere: release v1.2.56 (2026-11-13, atlas drawing-sheet skills, speech and audio and video parsing with duration billing and time-range citations); pushed 2026-11-14.
  • langchain: pushed 2026-11-14, 17,201 commits; langchain-core 1.8.0 still the latest tagged release.
  • lightrag: pushed 2026-11-14, 399 open issues aligned across sections.
  • llamaindex: pushed 2026-11-14; latest tagged release v0.14.26 (2026-11-13), the first since September.
  • open-codebase-index: side-branch push 2026-11-13 with the default branch still 2026-11-10; npm trailing month 5,055 to 2026-11-13, the 4,736-to-11-12 reading backfilled to 4,972, trend re-read roughly flat.
  • orama: rotation re-verification re-fetched every cited URL, all claims unchanged; npm trailing month 7,382,300 (window to 2026-11-13); name-squat weekly 246.
  • pageindex: PyPI SDK 0.6.0 (2026-11-14, 58 releases); pushed 2026-11-14.
  • r2r: rotation re-verification re-fetched every cited URL, all claims unchanged (pushed 2025-11-07, PyPI 3.6.6, docs 404, sciphi.ai unreachable, footprint unchanged).
  • unstructured: pushed 2026-11-14, 409 open issues; PyPI 0.27.37 still latest.
  • retrieval: matrix maintenance row refreshed to 2026-11-14 and the category index’s Orama one-liner aligned to about 7.4M installs; no membership change; front matter, links, and matrix sort, cell counts, and cell-to-note alignment verified clean across all sixteen members.
  • everos: recorded v1.7.0 (November 13, PyPI in sync at 1.7.0): the tracked pending line shipped (breaking everalgo 2026-11 contract with per-task LLM, profile items citing their source conversation, and the online purge API with a confirm step and job-status endpoint).
  • gbrain: recorded the train to v0.60.718.0 (10 releases, November 13-14): the push secret scan is now enforced with three bypass paths closed (managed-root scan-allow edits, clean-filter additions on strict canonical sources, git replace refs), the scan extended to escaped JSON and URL-encoded passwords, hard-deletes clearing six side tables, Cursor pinned-project skills, and the holding-page save performance fix.
  • openviking: recorded v0.5.5 (November 13, with CLI 0.5.3 and @openviking/react 0.1.11): the tracked November 13 batch shipped (pgvector vectordb backend, a PostgreSQL PathLock provider, the multi-node coordination setting without Redis, the snapshot-log path-history index, Console mode selection, and the Gateway Context Profile additions), with pi capture checkpoints and the privacy-config 409s still pending.
  • memu: recorded the pending November 14 work (local memory commits made atomic under concurrency).
  • screenpipe: recorded the pending November 13-14 batch (app v2.7.148 through v2.7.150, unreleased: configurable screenshot sampling with a 2 fps default, enterprise device policies without a webview with upload pins scoped to explicit MDM overrides, the macOS redaction capture-order fix, a Windows SQLite WAL-stall fix, and Claude login isolation).
  • supermemory: recorded the Pipecat voice-agent integration’s migration to the v5 namespace API (November 13).
  • zep: rotation re-verification, every cited URL live, the pricing ladder unchanged (monthly $125/$375 leading behind the Save-17-percent annual toggle, overage $25/$75 top-ups), and Graphiti’s commit count refreshed to 1,061.
  • cabinet: rotation re-verification, no facts changed (repository still quiet since August 25, pricing page unchanged, create-cabinet still 0.5.0).
  • claude-obsidian: rotation re-verification, no facts changed (v3.0.1 still latest, repository still quiet since October 25, no thread of its own).
  • mem0: rotation re-verification, no facts changed (BYOK pricing FAQ unchanged, default branch quiet since the November 6 security fixes, DolphinBench still dormant).
  • belief-context-graph: rotation re-verification, no facts changed (still zero tagged releases, default branch quiet since October 15, BrowseComp figures intact).
  • memory: memmachine, plur, zep, and belief-context-graph pushed_at signals resolved to side-branch pushes with default branches quiet since their last verifications, no note updates.
  • memory: matrix and category index verified matched to membership (34 columns and 34 entries, sorted, cell counts and alignment checked), no membership change this run.

(daily refresh, 3)
#

  • centaur: the calendar-versioning cut returned, v2026.11.13 (2026-11-13) topping the train above centaur-0.1.214; counters refreshed (292 forks, 121 open issues and pull requests)
  • claude-code-hooks: re-verified after the v2.1.325 release (2026-11-13), the 33-event table and all version gates still on the live docs, open issues roughly 17.2k
  • claude-code-action: added the February 2026 Stawinski writeup (external prompt injection via PR-title TOCTOU to workflow RCE, CVSS 7.7, fixed January 2026) as the critical source, found by the coverage scan; release train to v1.0.276
  • codex-action: fully re-verified in rotation with every cited URL re-fetched, all claims hold; counters 188 forks, default branch quiet since 2026-10-30 (side-branch push)
  • copilot-automations: fully re-verified in rotation with all five docs URLs re-fetched unchanged
  • n8n: counters refreshed (61,498 forks, pushed 2026-11-14)
  • executions: no credible dedicated source in the re-check; the category stays on the broad sources
  • valen: out-of-rotation board re-verification at the coordinator handoff, board data re-fetched identical to the 2026-11-13 read (generated 2026-11-12T12:34:44Z, 301 configurations), the two decision-model lists’ 2026-11-12 Valen commits resolved as catch-ups to the 2026-11-04 transfer (not a new URL move), the last pre-transfer-path citation re-pointed to the team org, and the release post’s vision-track readings for the 27B added (public Vision 82.9, full Vision 73.3 fifth, its 11 November snapshot)
  • jev: recorded Google DeepMind’s Gemini Decisions API entry (gemini-decision-1, announced 2026-11-11, the third hyperscaler on the noul/choice/score wire, $0.10 per million input tokens with output always free), with its first JevBench reading (thirteenth of 54 API rows by capability at 78.52, nineteenth on the composite)
  • decision-index: out-of-rotation re-fetch confirmed the board unchanged since the 2026-11-13 read, and the 2026-11-12 regeneration’s rank reads carried into the stale caution, bottom-line, and see-also mentions (Kev 27B seventy-fourth, the vendor’s row sixty-seventh)
  • jevbench: rotation re-verification, every cited URL re-fetched and live (repository unchanged since 2026-10-15, forks 33, thread steady at 154), and the boards rebuilt again (main board 261 ranked with Torchcast Decision 31B v1.0 first by capability at 89.45 ahead of new MetaDecision-31B-Rain and Karpo-Decision-31B, Torchcast Decision 4B still composite #1, API board 54 rows under a new composite leader decide-1.1 with gemini-decision-1 entering thirteenth by capability)
  • jevlike: rotation re-verification, all five references live, repository still frozen at 2026-09-16, nothing moved
  • semif: rotation re-verification, all seven references live, fork count 356 to 355 only
  • system1-arena: rotation re-verification, all five references live, headline rows unchanged, HN footprint still zero, dataset downloads 620 to 727 only
  • baml: volatile re-verification at the 2026-11-14 pushes, baml-wrapper 0.4.0 released 2026-11-13 with the 0.25.3 nightly train through the 20261112.a build, stable train holding at 0.25.2, npm trailing month 1,398,759
  • cua-s1: volatile re-verification at the host’s push to 2026-11-14, the component’s first change since the 2026-10-20 security fix being the 2026-11-13 pillow and soupsieve CVE bump, and the spaces v0.36.4-v0.36.12 and driver-rs v0.49.8-v0.49.9 releases not otherwise touching it
  • instructor: volatile re-verification at the push to 2026-11-14, two unreleased 1.18-line OpenAI performance commits, PyPI still 1.17.0, monthly downloads 9,829,816
  • strands-decider-2b: volatile re-verification at the push to 2026-11-14, the 2026-11-13 Lambda-images deploy feature on the default branch, the Decision Index row unchanged (20.72, rank 249 of 301), and the JevBench sibling row slipped to official #100
  • hybrid-execution: matrix maintenance cells moved for BAML, CUA-S1, Instructor, Strands Decider 2B, and JevBench, the JevBench provider-breadth cell to 261/54, the board paragraph re-read around both boards, and the category index one-liner moved to the new capability leader
  • code-review: open-code-review - volatile re-verification recorded v1.13.9 (November 13, run-coverage excluded files, zero-finding PR summary warnings, per-request tool declarations), 147 releases, pushed November 14.
  • code-review: roborev - volatile re-verification: push moved to November 14 with main active, v0.75.0 still latest.
  • code-review: openreview - rotation re-verification: dormancy re-confirmed (still pushed March 6, zero releases, no license file), no factual drift.
  • code-review: gito - rotation re-verification: every cited URL live, v4.5.0 still latest and pushed October 1, no factual drift.
  • code-review: code-review - q21 evidence: kodustech/awesome-ai-code-review still stale (pushed 2026-06-28, 54 stars), dedicated-list search surfaced nothing credible, the category stays on broad sources.
  • evaluation-review: crit - Recorded the v0.23.3 release (2026-11-13, two-way GitHub viewed-file sync, PR 1150) and moved the pushed date to 2026-11-13.
  • evaluation-review: hunk - Recorded the v0.25.0 stable release (2026-11-11, ends the 0.24 line, removes the deprecated –fast flag) and the 2026-11-13 extension-directory commit.
  • evaluation-review: langfuse - Recorded the v4.80.0 release (2026-11-13, ClickHouse request-body query parameters, Google AI Studio gateway provider) and refreshed the activity line.
  • evaluation-review: plannotator - Recorded the v0.28.64 release (2026-11-13) and the 2026-11-14 default-branch activity (Inbox bulk close, outside thread-id fix).
  • evaluation-review: pr-lens - Rotation re-verification: every cited URL re-fetched and live; the 2026-11-13 repository push identified as side-branch with the default branch at 2026-11-08; the stale seven-week age corrected to thirteen weeks.
  • evaluation-review: reviewbench - Recorded the twenty-fourth onboarding (Open Code Review, PR 110, merged 2026-11-14).
  • evaluation-review: verbatim-citation-gate - Recorded the v0.3.2 release (2026-11-14) shipping the four tracked false-found gate fixes and the ubuntu-24.04 CI pin.
  • evaluation-review: frontierharness-eval - Rotation re-verification: every cited URL re-fetched and live; recorded the live board’s v1.1 banner (21 harnesses, 5 models, 62 tasks) with the repository untouched since 2026-09-08.
  • evaluation-review: jevals - Rotation re-verification: every cited URL re-fetched and live; the JevBench board count refreshed from 235 to 261 ranked open-weights systems as of 2026-11-14.
  • evaluation-review: researchclawbench - Rotation re-verification: every cited URL re-fetched and live; the repository unchanged (last push 2026-10-28, the paper still at v5).
  • evaluation-review: memory-bench (LF-Labs-Inc/memory-bench) rejected on the zero-adoption floor with the vendor-self-scoring relationship recorded; reopen on any third-party signal or the q#85 ruling.
  • evaluation-review: valen - Out-of-rotation handoff re-verification: the Decision Index unchanged from its 2026-11-12 generation; the decision-model lists’ 2026-11-12 Valen commits identified as list-side URL catch-ups and the last old-path citation re-pointed to Valen-Team/Valen.
  • evaluation-review: phoenix - Volatile re-verification at the pushed signal: default branch through 2026-11-13 (daily dependency bump after the 21.1.0 release), the repository-level 2026-11-14 push side-branch; no note change needed.
  • e2b: recorded 2.64.0 (2026-11-13, template management on the SDK, resizable templates, OAuth access-token authentication, the commands.run shell option), PyPI at 2.64.0, pushed 2026-11-14.
  • openshell: recorded the CNCF Sandbox donation (2026-11-10, ai-openshell migration, Cedar policy language) in What-it-is, Status, and Cautions with the bet-on-the-vendor reading replaced, plus the post-v0.1.8 train through 2026-11-13 and open issues at 674.
  • container-use: the 2026-11-14 push checked as branch activity with the default branch still at August 2026, stalled-train reading stands.
  • artifact-fs: push advanced to 2026-11-13 on branch activity; Artifacts pricing page re-checked unchanged, still naming the now-passed 2026-11-12 billing start.
  • brig: 0.5.3 prerelease train through 2026-11-12, v0.5.2 still newest.
  • brood-box: push advanced to 2026-11-14 on branch activity, main still 2026-10-21, v0.0.25 latest.
  • microsandbox: post-v0.7.15 train unreleased, pushed 2026-11-14, issues 183.
  • nemoclaw: fix train through 2026-11-13, one-prerelease story holds, issues 869.
  • nono: v0.82.0 tag (2026-11-13) unpublished, v0.81.0 remains newest published release.
  • nvx: prerelease train advanced to 2026-11-14.
  • clawk: rotation re-verification, all cited URLs alive, ninety-three days quiet, reading hardened toward stalled.
  • greywall: rotation re-verification, all cited URLs alive, ninety-three days without a commit to main, v0.3.7 newest.
  • landrun: rotation re-verification, all cited URLs alive, v0.1.17 newest, no commit since 2026-10-16.
  • flue: rotation re-verification, all cited URLs alive, main-branch reading corrected to the 2026-10-12 v2.3.0 cut.
  • sandboxing-feature-matrix: E2B maturity to 2.64.x, OpenShell backing cell to the CNCF Sandbox donation, OpenShell policy cell records Cedar adoption.
  • sandboxing: category index and feature matrix re-verified matched to the 24-member category, sorted, cell counts aligned; all 24 notes’ front matter, links, and chart markup pass.
  • sandboxing: Agent Substrate admitted as the twenty-fifth member (Google-led agent execution runtime, CNCF Sandbox accepted 2026-10-20 and announced 2026-11-10, 1.0 GA with the managed GKE surface, gVisor or Cloud Hypervisor microVM isolation, actors snapshot-restored onto warm worker pools in under 500 ms); note written with fifteen fetched sources, the star-history chart at the dated baseline, the matrix column inserted in sorted position, and the category index entry added.

(daily refresh, 4)
#

  • spec-driven-development: ai-dlc - volatile refresh: pushed 2026-11-14 with 320 open issues, the stable train still v2.11.4 beside a v2.11.5 preview cut.
  • spec-driven-development: bmad-method - volatile refresh: pushed 2026-11-13 with 99 open issues, v6.12.4 still newest and the README’s mid-November v7 ETA passing unreleased.
  • spec-driven-development: gsd - volatile refresh: pushed 2026-11-14 with v1.21.0 still newest, npm figures 69,879 and 73,772 last month.
  • spec-driven-development: mycelium - recorded the v0.337.1 release (2026-11-13: correction counters aligned across session start, session end, and reflexion reconciliation, dated corrections answer reflexions with the ledger naming its counter baseline, stale pages fixed).
  • spec-driven-development: openspec - volatile refresh: pushed 2026-11-14 with the default branch at 2026-11-13, 206 open issues, npm 4,522,085 last month.
  • spec-driven-development: pdd-kit - volatile refresh: pushed_at 2026-11-14 on side branches with the default branch still at the v0.40.0 release commit, npm 2,134 last month.
  • spec-driven-development: spec-kitty - volatile refresh: pushed 2026-11-14 with v4.0.0rc8 still newest, PyPI trailing month moving 3,251 to 2,758 after ten identical reads.
  • spec-driven-development: spec-driven-develop - rotation re-verification of every cited source (all alive, repository quiet since 2026-07-26, still no HN thread); no facts changed.
  • spec-driven-development: tessl - rotation re-verification of every cited source (all alive, tiers unchanged at Free $0 and Team $100, CLI still quiet since 2026-03-05); no facts changed.
  • spec-driven-development: vibe-coding-prompt-template - rotation re-verification of every cited source (all alive, repository quiet since 2026-10-04); npm CLI figure moved to 170 last month.
  • spec-driven-development: matrix adoption row refreshed to 2026-11-14 as-of dates (GSD 70k npm, OpenSpec 4.5M npm, PDD Kit 2,100 npm, Spec Kitty 2.8k PyPI).
  • agentrq: refreshed counts (3 open issues and pull requests, pushed 2026-11-14); v0.15.7 still the newest release.
  • beads: refreshed counts (1,944 forks, 1,791 open issues, pushed 2026-11-14, 44,074 npm downloads over the trailing month to 2026-11-14); v1.3.4 still the newest stable release and rc.2 still the newest candidate.
  • epiq: refreshed counts (0 open issues and pull requests, pushed 2026-11-13, 3,227 npm downloads over the trailing month to 2026-11-13); v1.17.1 still the newest release.
  • its-a-plan: refreshed counts (308 forks, pushed 2026-11-14, 1,752 runner npm downloads over the trailing month to 2026-11-14); v1.10.0 still the newest release.
  • kata: refreshed counts (125 open issues and pull requests, pushed 2026-11-14); v0.23.0 still the newest release.
  • ordewell: refreshed counts (39 open issues and pull requests, pushed 2026-11-13, about 9,000 npm downloads over the trailing month to 2026-11-13); v0.13.1 still the newest release.
  • pad: refreshed counts (22 open issues and pull requests, pushed 2026-11-14); v0.20.0 still the newest release.
  • ai-dev-tasks: rotation re-verification, every cited URL re-fetched live (repository, API, both prompt files unchanged at 81 and 70 lines, the podcast link through its consent shell, the HN story search still at zero hits); still dormant with the last push at 2025-11-05, counts refreshed (1,697 forks, 216 watchers).
  • kanban-md: rotation re-verification, every cited URL re-fetched live (repository README with the config schema at version 12, the bundled development-loop skill, API, releases, the HN search with the launch thread still at 19 points); nothing changed, v0.39.0 still the newest tag, the last push still 2026-10-05.
  • automaker: rotation re-verification, every cited URL re-fetched live (repository page, README, LICENSE declaration, API, releases, the dead product site still 404, the HN search still car-industry only, the course site live, Anthropic’s credit-pool pause still standing); nothing moved (pushed still 2026-05-22, v1.0.0 still the final release, 584 forks).
  • task-management: matrix release-and-activity cells refreshed for the seven pushed members; the ai-dev-tasks, automaker, and kanban-md columns re-verified unchanged in rotation; no membership change; links, front matter, sort order, and cell-count parity verified across all 18 notes.
  • task-management: the dedicated-list check found no credible task-management-specific list for the fifth consecutive run; the TensorBlock awesome-mcp-servers Project-and-Task-Management docs section is an MCP-server list section, not a dedicated task list.
  • task-management: q24 evidence refreshed 2026-11-14: TomzxCode/agentic at 1 star, 0 forks, 1 open issue, pushed 2026-11-12, the coverage search returning only the owner’s own sites; the rejection stands, the owner-tools exception question stays open.
  • arifos: self-measurement stamp moved to 2026-11-14 (10 of 13 floors measured pass, 3 unmeasured, none failing, kernel degraded, vault999 stale, drift aligned) against the unchanged 2026-10-11 independent audit; PyPI caught up at 1!2026.11.13.post1 after holding at 1!2026.10.17 for a month, tags reaching v2026.11.14; counters refreshed (6 open issues, pushed 2026-11-14).
  • speakeasy-gram: release train moved through [email protected] to 0.3.52 and [email protected] (2026-11-13, MCP connects now require an explicit mcp:connect grant decoupled from mcp:read and mcp:write, and Anthropic inference hooks now scan tool call frames under the same policies as prompts) to [email protected] (2026-11-14); counters refreshed (283 open issues, pushed 2026-11-14).
  • vestige: release train moved to v6.2.2 (2026-11-12, the Forensic-8 zero-vector recall engine with self-measured RULER, LoCoMo, and LongMemEval-S figures); counters refreshed (234 open issues and pull requests, pushed 2026-11-14).
  • settlebridge: main’s commit count moved from 178 to 272 on a flag-gated fix-verifier wave (verification and warranty tables, payout gates on verified receipts with a 20 percent holdback, bypass-bounty submissions, GitHub role separation); counters refreshed (2 open issues, pushed 2026-11-11).
  • paperclip: backlog fell to 6,643 with the push date at 2026-11-14 and the homepage now advertising v2026.1106.0; added the release content from the Awesome Multi-Agent Orchestrators 2026-11-13 catch-up (Runner by default, Copilot harness, repositories as pinned skill sources, bring-your-own model gateways, model-backed prompt injection check); release line holding at v2026.1106.0.
  • agentfield: candidate line ran to v0.1.146-rc.31 (2026-11-13), PyPI holding at 0.1.145; counters refreshed (38 open issues, pushed 2026-11-14).
  • hol-guard: full rotation re-verification with drift: the coverage-contract page now Cloudflare-blocks fetchers while the product page on the same domain serves, so its renewal state stands as last read 2026-11-11 at 36 days past expiry; the product-page tracker moved to 2.7M+ downloads and 6.6K+ stars with a new 37M+ transactions counter while pypistats re-read 28,709 for the month as of 2026-11-14; counters refreshed (77 open issues and pull requests, pushed 2026-11-14); v3.59.0 holds.
  • okto-pulse: full rotation re-verification, every cited URL re-fetched, no drift (product page v0.4.2, 17 gates, 280+ tools; Neuron 0.3.1; nexus PyPI 0.3.2); counters refreshed (17 open issues, the 2026-11-13 repo push a side branch).
  • crowdstrike-falcon-guardian: full rotation re-verification, every reference re-fetched, no drift (docs path still a client-rendered shell, Reddit still 403).
  • microsoft-agent-governance-toolkit: full rotation re-verification, every reference re-fetched, no drift (launch post and the April 2026 critique unchanged, no release since v4.1.0, PyPI at 4.1.0); counters refreshed (151 open issues, pushed 2026-11-14).
  • agent-barn: counters refreshed (19 open issues, the 2026-11-14 repo push a side branch, default branch on 2026-11-11); v0.22.0 holds.
  • cc-safety-net: npm train holding at 2.7.12; counters refreshed (11 open issues and pull requests, pushed 2026-11-14 with the default branch on 2026-11-12).
  • cordum: counters refreshed (16 open issues, default-branch push 2026-11-13); v1.3.1 holds.
  • ibm-contextforge: counters refreshed (986 open issues, pushed 2026-11-14); v1.1.1 holds on PyPI.
  • mission-control: counters refreshed (61 open issues); v2.3.0 holds.
  • openappa: release train holding at v0.42.0; counters refreshed (36 open issues and pull requests, pushed 2026-11-14).
  • tracekit: release train holding at v1.6.0; counters refreshed (71 open issues and pull requests, pushed 2026-11-14).
  • control-planes: matrix volatile status cells refreshed for sixteen members plus arifOS’s governance cell; no membership change.
  • control-planes: the awesome-ai-governance 2026-11-12 to 11-14 additions resolve with one coverage-quality hold and seven rejections on standing families (see Entrant resolutions); the source’s 11-11 entries carry their 2026-11-11 resolution over, and the keel-verifier and Provenza commits edit already-resolved entries.
  • agent-analytics: rotation re-verification, all cited URLs live; the quiet stretch since 2026-10-09 ended (default-branch commits 2026-11-13 and 11-14: read-only POST /sql with property group-by, synchronous subject erasure for the self-hosted server; npm CLI 0.5.37 to 0.5.42 across 2026-11-12 to 11-14); the LICENSE file the README links still absent from the default branch and the Hacker News footprint still zero.
  • agentsview: volatile re-verification, no release after v0.47.0 with unreleased main-branch work headlined by a Direct API sessions parser for tools that call model APIs directly; the docs’ token-usage JSON contract flipped back to schema version 5 (the twelfth recorded flip) and the caution moved with it; forks 685 to 684, open issues and PRs 139 to 144.
  • ccusage: volatile re-verification, no release after v20.0.50 and npm latest unchanged, the pushes being routine pricing-snapshot, dependency, and Rust-1.100 build updates; trailing-month downloads refreshed (771,196 to 858,786, window 2026-10-15 to 2026-11-14 with five zero days, claude-mem at 159,572 so the multiplier moved to roughly five times), forks 919 to 922.
  • coucou: recorded Coucou for Android 0.5.0 (2026-11-13, another pairing-flow release note with no feature list) beside the unreleased Windows and Linux relay-settings merge; forks 1,174 to 1,181, open issues 460 to 451, pushed 2026-11-13.
  • git-ai: volatile re-verification, no release after v1.7.6 and no default-branch commit since 2026-11-05, the repository-level 2026-11-13 push landing on side branches; forks 327 to 330, open issues 258 to 265.
  • novafabric: volatile re-verification, no release after v0.108.0 and the default branch quiet since 2026-10-29, the 11-13 and 11-14 pushes landing on the dependabot branches (eleven open PRs, no open issues); PyPI still 0.108.0, trailing-month downloads 734 to 897.
  • codexmonitor: rotation re-verification, every cited URL live; dormancy unchanged (last commit 2026-03-26, v0.7.67 still the last release with the 0.7.68 bump unreleased) and moved to nearly eight months; forks 417 to 418, open issues 94 to 95.
  • clawtrace: rotation re-verification, all eight cited URLs live; repository state, npm plugin (0.1.23), credit packages and rates, and the paper all unchanged.
  • token-monitor: recorded v0.92.0 (2026-11-13, multiple Copilot accounts, featured-plugin auto-updates with a Trust-and-update review for third-party sources, quota refresh buttons, DeepSeek fee readout, launch-time usage with archived History) with unreleased work continuing (Qoder international usage, contribution guidance and a security policy, window and sync-status fixes); README headline re-verified 54-plus counting the aggregate; forks 355 to 358, open issues 149 to 143.
  • tokscale: recorded v4.22.1 (2026-11-13, archived OpenClaw SQLite usage dedup, Antigravity quota recovery across resumed sessions, Cursor refresh-grant isolation with OAuth redaction, invalid TUI color overrides ignored), npm latest 4.22.1; trailing-month downloads refreshed (123,064 to 125,330, five zero days), forks 476 to 475, open issues 77 to 71.
  • session-analytics: matrix volatile cells and prose moved to the same-day member facts (agentsview provenance shipped in v0.47.0, ccusage 19 named sources, Token Monitor 54 named tools, Tokscale 63 named clients, schema line back to version 5 as the twelfth flip, CodexMonitor choosing line to nearly eight months without a commit).
  • session-analytics: category index one-liners moved with the ccusage, Token Monitor, and Tokscale coverage counts.
  • fabro: volatile lines refreshed (pushed 2026-11-13, nightly train at v0.415.0-nightly.0); no material change
  • har: burst reached v1.25.1, twelve releases 2026-11-09 to 2026-11-13: portal-side attribution (one org connection v1.23.0, evidence-timeline shared v1.24.0, attributed usage shipped v1.25.0), out-of-order slot-evidence fix and Claude Opus 5.5 pricing v1.25.1; HAR HQ pricing unchanged
  • loki-mode: train at v11.140.0 (2026-11-13): base-red suite excuse on identical head failure, per-package test runners, plan-scope intake guard; Docker pulls 326,874
  • machinist: main tip still 2026-10-29 with the 2026-11-09 push a side branch, v0.4.0 still newest, 179 commits past; zero coverage holds
  • open-inspect: pushed 2026-11-14; release list still empty; zero coverage holds
  • open-swe: rotation re-verified; newest stable desktop tag Robin Desktop v0.3.16 (2026-11-09), which the releases list buries under its nightly flood
  • superplane: pushed 2026-11-14; newest main commit a merge-confidence running check (2026-11-13, unreleased); no release past fleet-manager v0.1.7
  • zeroshot: v16.0.0 (2026-11-13): in-place delivery the breaking default with –worktree opt-in isolation; five interface-breaking majors v12.0.0 through v16.0.0
  • super-simple-software-factory: rotation re-verified; still one commit on main, no push since 2026-08-04, no releases
  • software-factory: matrix caught up to its notes (Ouroboros 15 runtimes, SuperPlane repriced plans) and moved the Zeroshot v16.0.0 isolation cell; varun1505 2026-11-09 and 2026-11-12 sweeps resolved with no admissions, Pullboard 820 held below the bar on the Orbi coverage precedent

(daily refresh, 5)
#

  • kimi-code: recorded the K3 OpenRouter listing’s twenty-second move (2026-11-14): main listing held at $0.192/$16.00 (cache reads $0.182) while the input floor rose on Relace’s fp4 repricing to $0.1824/$11.0055 (cache reads $0.18), Wafer’s priority fast row cut to $0.153/$16.00 (cache reads $0.152) as the cheapest visible input, and the Decart and Morph rows left the roster (30 to 29 endpoints) with the output floor holding at Chutes $1.85/$9.50; Moonshot’s own kimi-k3.1-pro rows held $2/$10 (highspeed $4/$20) and kimi-k2.7-code’s rows carried its recorded $0.60/$3.50 (cache reads $0.16) reading.
  • openrouter: pricing page re-fetched on the daily-churn list with the fee state unchanged (5.5% Standard, 8% Business, BYOK free to $25,000 then 5%, workspace 5/5/1,000/custom), as-of moved to 2026-11-14.
  • experiential: recorded the exp-code line shipping 0.1.3 through 0.1.7 across 2026-11-13 and PyPI advancing to 0.7.242 at 247 releases (the 11-14 release-bump commit on main unreleased on GitHub past v0.7.234); repo pushed 2026-11-14.
  • litellm: recorded v1.108.7 (2026-11-12), v1.111.0-dev.1 and v1.111.0-dev.2 (2026-11-13) taking the train head past the never-stable v1.110 line, v1.108.8 (2026-11-13), and v1.106.10 (2026-11-14); PyPI latest holding at 1.109.1, repo pushed 2026-11-14.
  • bifrost: pushed date refreshed to 2026-11-14 (default-branch cache-breakpoint fix), no release past the 2026-11-10 patch train.
  • llm-gateway: pushed date refreshed to 2026-11-14 (model-page UI work), v1.25.0 still latest.
  • ollama: pushed date refreshed to 2026-11-14, v0.40.16 still latest with v0.40.17-rc0 on the train.
  • localai: pushed date refreshed to 2026-11-14 (Explore-page UI work), v4.12.1 still latest.
  • magnitude: pushed date refreshed to 2026-11-13 (default-branch tooling fix), cli 0.2.17 still latest.
  • model-access: delta-pass verification clean across all 26 notes, the matrix, and the category index (zero broken internal links over 28 files, no type fields, matrix rows at the header’s 27 cells, membership and case-insensitive sort unchanged, no cell moves from the release and push activity, category fit unchanged).
  • model-access: the Cursor shutoff-date handoff resolved with nothing to correct: no model-access note and neither matrix cites the November 12 date, all in-scope Cursor mentions are compatibility or SuperGrok-bundling facts, and the date-corrected reading lives only in the surfaces note and surfaces index.
  • benchmark-radar: volatile re-verification after the 2026-11-14 pushes: the Hugging Face dataset card’s license switched to MIT (PR #1256, matching the repo’s data-files grant), the site gained a Talks page with the Cohere Labs Open Science Community talk, Discord and WeChat group surfaces joined the README, badge to 49,679 records over 109 days; v0.14.0 still newest release, HN footprint still zero.
  • bestvaluemodel: volatile re-verification after the mid-November pushes: the 2026-11-09 run added nine rows headed by all five Claude Fable 5.2 settings (Max at 64.6, the page’s new top score, at a blended $16), the 2026-11-12 run added the four Gemini 4 Hydrogen settings and halved the Argon and Carbon price lines ($8 to $4 blended, reversing the 10-28 doubling), the 2026-11-14 run added Gemini 4 Argon (Medium) and DeepSeek V4 Pro 0813 (High); repo still unlicensed, bot cadence unchanged.
  • ai-release-tracker: rotation re-verification with every cited URL re-fetched: the archived homepage’s newest capture moved to 2026-11-02 (19 captures), reading 270 models from 11 companies with a newest tracked release of Gemini 4 Carbon (October 28), the $499/month slot, To sider ApS footer, and expected-next cards unchanged, the operator’s xAI spelling now reading SpaceXSI, per-model pages and the sponsorship contact captured the same day, the corpus flat files still captured only 2026-08-25/26 with the live fetch still 429, and the BI citation, Show HN thread, and single-story footprint re-verified unchanged.
  • whatmodel-sh: rotation re-verification with every cited URL re-fetched: coverage to 891 ranked text models from 37,490 published benchmark results across 575 benchmarks (board stamp 2026-11-03) with a new /decision list, the 2026-10-30 terms revision prices the sponsor placements from US$99/month through Stripe (note gains its Price history table), the about-the-author card names creator Jake, repository pushed 2026-11-03, HN footprint still one story.
  • trackers-and-leaderboards-feature-matrix: cells refreshed from the run’s volatile re-verifications (Benchmark Radar coverage and licensing, bestvaluemodel cadence unchanged) and rotation re-verifications (AI Release Tracker coverage and cadence, whatmodel.sh coverage, cadence, pricing, and run-by); membership unchanged.
  • trackers-and-leaderboards: the dedicated-awesome-list check re-ran 2026-11-14 and still finds no credible dedicated source.
  • navan-chauhan: created; admitted on the StrongDM factory reveal (304 HN points, in-window breakout, Willison same-day coverage) with his own factory essays as the sustained output; matrix column, index entry, and sort landed same run.
  • owain-lewis: rotation re-verified, two new uploads (2026-11-02, 2026-11-11) make 21 between May 15 and November 11, newest video 30,461 views in three days, funnel adds a free Skool community alongside the paid one.
  • paul-gauthier: rotation re-verified, every cited URL live, stall holds (default branch quiet since 2026-05-22, no release since v0.86.0, leaderboard frozen at October 2025).
  • peter-steinberger: rotation re-verified, blog moved to steipete.md with steipete.me redirecting every page, citations moved to the canonical domain, 1,449-point announcement thread re-grounded through its direct Algolia item after the loose query decayed.
  • shreya-shankar: rotation re-verified, DAB leaderboard advanced to entries dated through 2026-11-02 and the repo to 207 stars (last push 2026-11-07), blog still dormant since May 21.
  • simon-willison: rotation re-verified, daily posts through 13 November 2026 led by the Datasette plugin releases and the llm 0.37.2 and llm-gemini 0.37 releases.
  • people-and-publications: Patrick Debois rejects as a voice on the 2026-10-13 long-established clause (DevOps-era breakout predates the window); his 2026 agent-era stream on jedi.be and tessl.io with Camplight independent coverage recorded as the reopen condition.
  • people-and-publications: Eno Reyes (Factory CTO) rejects as a voice pending q10, vendor-executive output living on company channels matching the never-admitted Zach Lloyd profile.
  • people-and-publications: Osmani’s Code Agent Orchestra essay (2026-03-26) and the Latent Space Warp episodes (2026-07-01 written, 2026-11-03 podcast) verified as member own output in scope, no note change.
  • people-and-publications: matrix extended to thirty-four columns with Navan Chauhan; Owain Lewis cadence and Steinberger platform cells moved on this run’s verifications.
  • section-root: model-provider-feature-matrix: flagship cells moved to Claude Fable 5.2 (2026-11-09, $8/$40, cache $0.20) and Gemini 4 Carbon (2026-11-10, Argon’s ladder), Hydrogen ($1.50/$7.50, 2026-11-12) folded into the workhorse and cache cells, Kimi floors re-split (Wafer $0.153/$16.00 input, output back to Chutes at $1.85/$9.50 after Relace’s output re-priced to $11.0055, roster 29), kimi-k2.7-code floor held at $0.60/$3.00, models.dev’s kimi-k3.1-pro weights read back to Open.
  • section-root: model-selection-for-coding-tasks: Fable 5.2 column added, Google Carbon and Hydrogen recorded, gpt-6-cyber noted, Kimi floors re-dated, harness-fit paragraph rewritten for OpenCode’s models.dev-backed /connect catalog (the providers docs no longer name vendors), stale 2026-11-21 promo bullet corrected to 2026-11-30.
  • section-root: agentic-coding-tools-landscape: Cursor shutoff date-corrected (temporary extension past November 12, models docs still list Sol/Luna/Terra with no removal notice, CNBC replaces devops.com), DeepSeek Harness rc.3 (November 13), Gemini CLI docs stamp 2026-11-10, Cline about 5.9 million installs, Cursor pricing re-checked unchanged.
  • section-root: model-benchmark-matrix: Terminal-Bench 4.1 per its homepage.
  • section-root: the-agentic-development-environment-landscape: rotation pass clean (17 of 17 URLs 200), Air’s page now also names Codex, Junie, Copilot, and OpenCode, Paradev MIT and Paseo wrapper re-confirmed, Agent Orchestrator’s own description reads 35+.
  • section-root: context-management-patterns: link check clean, compaction claims hold, as-of holds at 2026-11-03.
  • section-root: the-tells-are-structural: sources re-fetched, figures and ruleset ids hold, as-of holds at 2026-11-03.
  • kapso: recorded the v0.9.2 and v0.9.3 releases (November 11-12): kapso compare, the minimize-direction fidelity fix, the agent-engineering and data-engineering harnesses bundling into the package, setup-only task examples, and the post-release report-json and walkthrough-video commits.
  • openresearch: recorded the v0.2.27 release (November 13), which shipped the commits the November 11 refresh had read as post-release work, the November 13-14 post-release train, and the open-issue count moved to 42.
  • local-deep-research: recorded the November 11-14 post-release batch (2.0.0 upgrade safety, configurable sampling parameters, parser temp-dir isolation, publisher-PDF discovery, empty-database refusal) and the open-issue count moved to 1,494.
  • autosci: pushed 2026-11-13, another WeChat-group QR refresh.
  • anthropic-claude-math: rotation re-verification, no drift, formal-math default branch quiet since October 27.
  • autoresearchclaw: rotation re-verification, no drift, dormant since 2026-08-19.
  • evo: rotation re-verification, no drift, quiet since 2026-10-05.
  • sakana-ai-scientist: rotation re-verification, no drift, v1/v2 dormant since 2025-12-19 and v3 quiet since October 18.
  • aiscientist: rotation re-verification, no drift, dormant since July 2026.
  • automated-research: matrix Kapso and OpenResearch surface cells moved to v0.9.3 and v0.2.27.
  • kimi-code (harnesses): mirrored the model-access pass’s twenty-second endpoint move into the harnesses note (roster 29 after Decart and Morph left, the Relace fp4 repricing lifting the standard input floor to $0.1824/$11.0055 with Wafer’s priority fast row at $0.153/$16.00 cheapest visible input, main listing and Chutes output floor held).
  • agent-orchestrator and its matrix cells: coordinator correction from the section-root pass, the supervised-CLI count moved from 26-plus to 35-plus on the repository’s own description (fetched 2026-11-14).
  • cursor: coordinator correction from the section-root pass, the GPT-5.6 Sol promotional-pricing callout is back on the fetched models-and-pricing pages after the 2026-11-08 pass recorded it gone; the Status line carries both readings dated.
  • agents/_index.md: essay and matrix one-liner verification dates bumped to 2026-11-14 for the five section-root articles refreshed today.