## (daily refresh, 1)

- cross-category: stars scan clean, 2 new stars since checkpoint (spiffe/spire, HewlettPackard/py-spiffe), both rejected on the no-agent-story scope line; checkpoint advanced to 2026-11-06T07:28:48Z
- cross-category: pandacode (aicodeking/pandacode, 1,109 stars as of 2026-11-07) admission resolved for orchestration on the standalone-ADE line, above the 751-star 2026-11-01 wave floor, MindStudio and cnblogs third-party writeups, PandaCode-License and unsigned-installer cautions recorded
- cross-category: LazyMind (LazyAGI/LazyMind, 612 stars) held out-of-scope pending an owner ruling, no claiming category after kind declines (skills runtimes excluded, not multi-agent orchestration, plugin-into-harness not harness-kind)
- cross-category: caprock (dspv/caprock, 41 stars), devthrottle (110), antumi (97), claudette (87), and prometeu (64) rejected on the thin-record below-band line while the tier question is open
- cross-category: andyrewlee and AMO additions scanned, deja-vu and Orchestrator.inc already-covered listings, Happier/Agentbox/Agetor re-additions carry; the andyrewlee Resting sweep moved AgentsMesh to dead, flagged to orchestration
- cross-category: ai-for-developers, hesreallyhim, e2b-dev, and slavakurilyak carried over unchanged (newest commits 2026-09-18, ticker-only, 2026-08-21, 2026-10-04 ticker)
- model-provider-feature-matrix: delta-pass price check, the Moonshot flagship cell's OpenRouter floors re-converged on Relace's re-priced fp4 row at $0.1824/$9.3388 holding both legs, Wafer's fast row re-priced up to $0.192/$16.00, Chutes' mxfp4 row held at $1.85/$9.50 without either floor, the aggregator listing holding at $0.192/$16.00, the roster narrowing to 30, and the kimi-k2.7-code floor held at $0.60/$3.00 on CoreWeave's int4 row
- model-provider-feature-matrix: all seven provider pricing pages re-fetched with every other vendor cell unchanged, and models.dev's canonical Weights column read K3.1 Pro Closed against Moonshot's open-weight card and the public Hugging Face release, recorded as a reference disagreement with the weights cell grounded on the primaries
- model-selection-for-coding-tasks: synced the K3 floor reading, the decision-guide and whole-repo bullets, and the fourth-clock paragraph to the 2026-11-07 Relace-converged readings, gpt-5.6-terra and gpt-5.6-luna still listed with Sol's promo through November 30
- model-benchmark-matrix: link re-verification, every cited URL 200 except the OpenAI PaperBench page again 403 to the scripted fetch, Terminal-Bench still at 4.0, no fact moved
- agentic-coding-tools-landscape: link re-verification, sixteen of eighteen fetched 200 with Bloomberg and devops.com again 403 and both GitHub topic pages 200, DeepSeek Harness newest prerelease unchanged at dsh-v0.2.2-rc.1, the Gemini CLI enterprise-docs banner unchanged, and the Cursor X-bundle plans unchanged, no fact moved
- the-agentic-development-environment-landscape: rotation re-verification, all 17 URLs 200, Air's supported-agent, surface, and cloud-run claims, Paradev's MIT license, Paseo's Apache-2.0 wrapper, and Agent Orchestrator's 36-entry supported-agent table all re-confirmed, no fact moved
- context-management-patterns: link check, all nine cited pages 200 with the compaction-survival, skill-cap, idle-compaction, v2.1.304, and /autocompact claims re-checked against the live docs with nothing to correct, no fact moved
- the-tells-are-structural: re-fetched all eight cited sources, every one 200, the StoryScope and human-detector figures re-checked unchanged and the no-ai-slop ruleset's structural rule ids intact, no fact moved
- section-root: section index stamps synced mechanically from each category matrix's updated date, header row, and Changes tail, closing the gap the 2026-11-06 matrix passes left (ten stale counts, dates, and newest-column claims across automated-research, code-review, control-planes, evaluation-review, executions, memory, model-access, orchestration, protocols, retrieval, and surfaces, plus the three essay date stamps)
- orchestration: kaushikb11/awesome-llm-agents unchanged since 2026-11-06, entrant scan carried over with no new candidates
- orchestration: category checks clean (links, front matter, matrix membership and sort, index list and sort, chart markup) across all 89 members; no membership change
- orchestration: 18 notes' Changes sections re-sorted oldest-first (a past run inserted the 2026-11-06 bullet mid-history), content untouched; ag2, bernstein, langgraph, and omnara were sort-only repairs with no updated bump
- agent-manager: recorded the repository transfer from YoanWai to the wainmann account (old path a 301); chart and references moved to canonical wainmann/agent-manager paths; rotation re-verified with all cited URLs live
- squad: recorded the repository transfer from bradygaster's account into the github organization (old path a 301); chart and references moved to canonical github/squad paths; push refreshed to 2026-11-06, v1.1.0 still latest
- vibe-kanban: recorded the 2026-11-03 community commit (PR #3490), the first since 2026-10-28; fifteen community commits since the shutdown; dropped the superseded only-activity-in-the-month claim
- agent-orchestrator: recorded v0.14.10 (2026-11-06) as the latest stable release and the 2026-11-07 push
- agent-swarm: recorded v2.8.3 (2026-11-07) at the head of the v2 train and the 2026-11-07 push
- agentlas-os: recorded v1.2.88 (2026-11-07) as the latest release and the 2026-11-07 push
- crewai: recorded 1.15.37 on GitHub and PyPI (2026-11-07) and the 2026-11-07 push
- emdash: recorded v1.2.11 (2026-11-06) as the newest release; rotation re-verified, all cited URLs live
- graphcode: recorded v0.1.84 (2026-11-06) as the latest stable release and the 2026-11-07 push
- jean: recorded v1.0.14 (2026-11-06) as the latest release and the 2026-11-06 push
- lobehub: recorded v2.3.2 (2026-11-06) stable and the 2026-11-07 push
- looptroop: recorded v0.6.4 (2026-11-06) as the latest release and the 2026-11-07 push
- multica: recorded v0.8.1 (2026-11-06) as the latest release and the 2026-11-07 push
- nodeterm: recorded v0.5.1 (2026-11-06) and the 2026-11-07 push
- openharness: release lines refreshed across all three tracks (v0.5.66_cli 2026-11-07, v1.3.99_web 2026-11-06, v1.3.89_desktop 2026-11-07) and the 2026-11-07 push
- openrig: recorded v0.7.9 (2026-11-06) as the newest release
- orca: recorded v1.4.252 (2026-11-07) as the latest desktop release
- paseo: recorded v0.15.3 (2026-11-06) in the release train
- ruflo: recorded v3.72.0 and v3.72.2 (both 2026-11-06) closing the train, no v3.72.1 tag in the release list
- sim: recorded the release train through v0.9.47 (2026-11-07) and the 2026-11-07 push
- amp: rotation re-verification, every cited reference live, recorded the November train through Spaces (November 6), tier page unchanged
- claude-code: rotation re-verification, every cited reference live, recorded npm 2.1.315 through 2.1.319 (November 6 latest) with the stable dist-tag parked at 2.1.311
- command-code: rotation re-verification, every cited reference live, npm train to 1.88.5 (November 6), download window refreshed to 174,498, site changelog counter stalled at 468
- fx: rotation re-verification, every cited reference live, v0.0.16 still newest, custom-model-connections docs dropped their preview banner and the caution was rewritten
- grok-build: rotation re-verification, every cited reference live except the 403ing changelog page, stable promoted to v1.0.77 (November 5) with alpha at v1.0.78, download window to 309,565, API prices unchanged
- longhorizon-harness: rotation re-verification, every cited reference live, v0.1.8 still newest with no commits since 2026-10-21, pypistats re-read unchanged at 1,048
- zerostack: rotation re-verification, every cited reference live, crates.io counters to 3,783 total (690 recent), v1.8.5 still newest, push date unchanged at 2026-11-04
- junie: recorded the release train running on, eight installer builds published November 6 through 4040.1
- oh-my-pi: recorded npm 18.18.0 (published November 7, fix train), push date refreshed to 2026-11-07
- openhands: recorded the v1.37.0 release (published November 6), push date refreshed to 2026-11-07
- qwen-code: recorded v0.26.9 (published November 6), bulk skill activation over a new daemon API and disk-based ZIP handling with recovery backups
- opencode: recorded the v2 train reaching v2.2.6 (published November 6, shared MCP connections, GUI plugin themes), process-daemon train at v0.5.0
- copilot-cli: recorded npm promoting 1.0.107 to stable (published November 6, prerelease at 1.0.108-1), download window refreshed to 6,377,474
- gemini-cli: refreshed the commit count to about 6,600 (as of 2026-11-07)
- harnesses: push dates refreshed to 2026-11-07 on kilo-code, gptme, jcode, nanocoder, om-o, and omnigent
- harnesses: matrix and category index verified in sync with membership (45 members), front matter and internal links clean
- orchestration, harnesses: volatile push-date and as-of refreshes across the remaining flagged members (archon, autogpt, buzz, cezar, chatdev, cli-agent-orchestrator, cmux, commonly, dify, foremerge, gascity, happier, happy-coder, mastra, orkas, parallel-code, praisonai, sandcastle, scion, superset, t3code, tuios, worktrunk, apowerb, crewplane, and others), no content changes beyond the ones listed above

## (daily refresh, 2)

- pandacode: admitted to orchestration, 1,109 stars as of 2026-11-07 (400 on 2026-10-22), v0.3.101 on a daily release train, nineteen engines, MIT-plus-fork-attribution license reading NOASSERTION on the API, MindStudio 2026-10-21 and cnblogs 2026-10-22 writeups as the only third-party coverage, zero HN footprint, creator-benchmarks caution recorded; matrix column and index entry added.
- orchestration-feature-matrix: PandaCode column added (eighty-nine to ninety columns) with counts bumped and the Changes section re-sequenced (2026-11-03/04 bullets into order, two 2026-11-06 bullets re-housed from after References, stale eighty-three-notes count fixed).
- agentsmesh: awesome-agent-orchestrators moved the entry to Resting on 2026-11-05, recorded in the note as a quiet signal.
- agent-orchestrator: Orchestrator.inc product name recorded from the awesome-agent-orchestrators listing and the docs site.
- personal-agent-protocol: created from the coordinator's 2026-11-03 admission with the Status fact corrected from live fetches (draft 0.1 shipped 2026-10-09 as the Poppy draft, 0.2 on 10-24, 0.3 frozen 11-04, draft 0.4 open), 43 listed design partners, 77 stars, the one-of-44-partners deployment scan, and beri.net as the standing critical source; matrix column and index entry added.
- protocols-feature-matrix: extended to fourteen columns with Personal Agent Protocol (Poppy), inserted in sorted position between MCP and PIC Standard, with the intro, thesis, governance count, adoption, criticism, and choosing sections updated.
- ag-ui: train moved to release/2026-11-06 with the first ag-ui-openai-agents Python package (0.1.0), downloads refreshed to about 21.3M combined (month ending 2026-11-06)
- a2ui: pushed 2026-11-07 on docs-site edits and a widened React peer range, tagged train unchanged at the 2026-11-02 Python wave
- acp: default branch stabilized the client system prompt (2026-11-06, unreleased) with trains, SDKs, and the 49-entry agents list unchanged
- anp: rotation re-verification, no dead sources, hub still ANP 1.2, footprint unchanged, the 2026-10-30 push corrected to a side-branch push with the default branch quiet since 2026-10-22
- llms-txt: rotation re-verification, no dead sources, spec site still v2, Anthropic's served file now 964 lines, footprint unchanged
- protocols: NeuroCode (Arkz-Deepak/neurocode, 0 stars) rejected on the servers-and-utility-layers ruling, context-slicing kind flagged for context-engines routing; mcp-clients and awesome-agent-protocols scans carried over unchanged
- centaur: 2026-11-06 admission note verified out of rotation, every cited URL re-fetched; counters refreshed (284 forks, 120 open issues), release train moved to calendar versioning (v2026.11.6, 2026-11-06), contributed Teams, GitHub, Discord, and Linear ingress recorded; matrix column and category-index entry added with the Changes bullet dated to the admission
- executions-feature-matrix: added the Centaur column (twelve to thirteen members, sorted alphabetically), rebuilt the table with a separator row matched to the header's cell count, extended the Kind, trigger, open-source, guardrails, versioning, and choosing prose, and updated the intro counts
- airut: rotation re-verification, every cited URL re-fetched, all claims re-confirmed unchanged; counters refreshed (2 open issues, pushed 2026-11-07)
- run-gemini-cli: rotation re-verification, every cited URL re-fetched, stalled status re-confirmed (push still 2026-08-21, v0.1.22 still last release); counters refreshed (292 forks, 88 open issues)
- aeon: volatile refresh, counters (300 forks, 1,546 commits, pushed 2026-11-05) and proof-of-work counts (152 findings across 113 repositories)
- claude-code-action: volatile refresh, release train v1.0.269 (2026-11-06) and counters (2,254 forks, 858 open issues); Latest badge still on v1
- claude-code-hooks: volatile refresh after v2.1.319, the 33-event table and the v2.1.267/v2.1.295 gates still on the live docs, open issues roughly 16.8k
- github-agentic-workflows: volatile refresh, v1.2.1 prerelease (2026-11-06) tops the train, forks 619, README advisory notice still live, no new advisories
- rivet-agents: volatile refresh, v0.5.6 (2026-11-05, Pi 1.5 support, stable RivetKit 3, OTel tracing for Pi Durable) and counters refreshed
- trigger-dev: volatile refresh, counters (1,526 forks); v4.8.3 still newest, push still 2026-11-03
- code-graph-rag: refreshed activity (pushed 2026-11-07 on the version bump to 0.5.7, 520 open issues), re-fetched PyPI (0.5.0 still latest, 38 releases), corrected the 0.4-line endpoint to 0.4.19 (2026-11-02) and the stale PyPI sentence lead; the 2026-11-06 interrupted-run verification stands
- langchain: refreshed activity (17,153 commits, pushed 2026-11-07); latest tagged release langchain-core 1.7.6 unchanged
- ragflow: recorded v1.0.0 (2026-11-06), the GA cut of the Go rewrite following rc3, refreshed activity (pushed 2026-11-06, 1,333 open issues), noted the repository description's repositioning to "an agent-native database"
- unstructured: recorded PyPI 0.27.37 (2026-11-06, 249 releases) and refreshed activity (pushed 2026-11-06)
- pageindex: recorded the SDK train reaching 0.5.2 (2026-11-04) and 0.5.3 (2026-11-05, 57 releases) and refreshed activity (pushed 2026-11-07, 167 open issues)
- open-codebase-index: refreshed activity (repo-level push 2026-11-04 was a side-branch push, default branch latest commit 2026-11-01) and re-read trailing-month installs at 5,118 (seven zero days), roughly flat; npm train paused at 0.38.2
- llamaindex: refreshed activity (pushed 2026-11-06, 1,307 open issues counting pull requests)
- graphrag: rotation re-verification re-fetched every cited URL live; all claims unchanged (maintenance-mode declaration, v3.3.0, 52 releases, the 282-point launch thread)
- tree-sitter-chunking: rotation re-verification re-fetched every cited URL live; all claims unchanged
- semantic-code-search: interrupted-run 2026-11-06 verification kept; VS Code workspace-context page date moved to 2026-11-04 with claims unchanged
- retrieval: the seven 2026-11-06 interrupted-run verifications spot-checked and kept, closing chonkie, code-graph-rag, docling, knowhere, langchain, lightrag, and semantic-code-search rotation; code-graph-rag and langchain pushed again 2026-11-07 and were refreshed in the same pass
- retrieval-feature-matrix: refreshed the maintenance row to 2026-11-07 (RAGFlow v1.0.0 GA, Docling PyPI 2.148.0, Knowhere v1.2.45, PageIndex PyPI 0.5.3 pushed 2026-11-07, Unstructured PyPI 0.27.37, Code-Graph-RAG pushed 2026-11-07, open-codebase-index 5.1k installs); no membership change
- retrieval: both mapped sources unchanged since their 2026-11-06 checks, entrant scan carried over with no new candidates

## (daily refresh, 3)

- agentmemory - recorded v0.9.32: LLM and embedding providers off by default behind explicit environment variables, hooks refuse plain-HTTP remote servers, PreCompact injection opt-in, plus the shipped recall train, MCP tool annotations with a privacy policy, and Codex/Hermes/OpenClaw plugin listings
- claude-mem - recorded the v15.10 and v15.11 lines: onboarding discovery rework, capture kill switch, per-source observer providers, long-lived Claude login env token, installer setup-screen telemetry, and the stale-observer-work replay fix
- cognee - rotation re-verified every cited URL live, pricing unchanged; recorded v1.6.7: ingestion refuses raw database connection strings and HTTP server file paths by default, scoped searches corrected, query scoring batched
- deja - recorded v0.24.0: sub-50 ms hook and MCP hot path, project X-Ray and symbol-decision recall, PreCompact working-set capture, asymmetric Nemotron/NIM embeddings, native Antigravity plugin bundle
- funes - recorded v1.11.0: Lance index-refine vector search, index-refresh row fix, rmcp 3.8.0 / Lance 13.0.0 / hf-hub 1.0.0 bumps
- gbrain - recorded the train to v0.60.601.0: identifier and handle search, the memory train, HTTP MCP agent attribution, opt-in Wakeshift wakes, the shared-channel privacy fix, and the four-hole /cso-audit security wave
- letta - rotation re-verified every cited URL live, pricing ladder unchanged; letta-code's latest release refreshed to v0.34.44
- okf-agent-memory - recorded v0.8.0: declarative bundle scope isolation with fail-closed refusal, two-stage concurrent write protection, bootstrap SSoT symlinks, breaking --scope bundle alias removal
- operator-memory - recorded the November 5 releases: helper v1.12.0 and codex v0.4.1
- screenpipe - recorded app-v2.7.143 and v2.7.145: enterprise uploads fail closed without an upload destination pin, heartbeats report device timezone, fresh histories sync from persistent database sources
- sno-station - recorded the 2.0 line: v2.0.0 tag-team and v2.0.1 timeout resilience; v1.7.x still standing
- zep - recorded the repository's removal of the legacy Community Edition code and README section, with Agent Skills harnesses and seven-framework examples replacing it
- emulo, engrim, memanto: rotation re-verified, every cited URL live, no material change
- memory: the TeleAI list's 2026-10-31 to 11-06 additions resolved on standing families (seven below the 294-star floor, one thin record, two closed-hosted, one routed out on the primary-artifact reading, four paper entries, one benchmark routed to evaluation-review); Comma (AFK-surf/Comma, 514 stars) flagged to assistant-runtimes
- baml: stable train resumed, baml-language 0.24.0 released 2026-11-05 through 0.25.2 on 2026-11-06, the 0.23.1 nightlies ended, pushed 2026-11-07, npm trailing month 1,309,425
- cua-s1: host pushed 2026-11-07, the cua-s1 component untouched since the 2026-10-20 security fix
- instructor: pushed 2026-11-06 on the unreleased 1.18 line, PyPI 1.17.0 unchanged, monthly downloads flat for a sixth check
- jeff: repositioned into a decision-function library with a 34-adapter catalogue, v1.4 base released 2026-11-05, five unpublished SDKs, and a hosted API at api.jeffhub.ai, pushed 2026-11-06 after thirty days quiet; README benchmark table dropped for a dated 2026-10-17 head-to-head
- laya: 0.4.25 through 0.4.27 released across 2026-11-05/06, PyPI 0.4.27, pushed 2026-11-06
- ollaya: pushed 2026-11-06 on three unreleased commits, v0.18.0 still the ceiling, site at thirty families
- strands-decider-2b: v0.3.0 released 2026-11-05 (PyPI 0.3.0, the consolidated serving-and-deployment release), pushed 2026-11-06, Decision Index row 225 of 275 at unchanged 20.72
- decision-index: the 2026-11-06 regeneration grew the board to 275 configurations with REAX 1.3 first at 73.98, the vendor's rows at fifty-fifth and sixty-fourth, the vision board at 78 entrants
- h2o-lightning-4b: rotation re-verification, the card's 2026-11-01 rewrite documents v1.4.2 and adds ImageJevBench claims, the v1.1 row holds official composite #2 at 77.94 with wald-26b-a4b newly third
- jevbench: board re-fetch stable at 235 ranked and 48 API rows with the composite top two unchanged and wald-26b-a4b newly third
- jev, kev, nimble, von, winnow, anyjev, clm, decider: board re-fetches at the Decision Index's crowd-shift, scores unchanged, ranks moved; nimble gained a v4.1 Max row at 70.03, eighth of 275
- hybrid-execution: feature-matrix maintenance cells moved for the seven pushed or released members and both boards' cells moved to their current states, no membership change
- hybrid-execution: the category's first credible dedicated list appeared, andyrewlee/awesome-system-one, wiring proposed in the queue
- bifrost - release train refresh: transports v2.5.0 landed 2026-11-06 with the MCP-gateway wave plus plugin patches on core v1.15.1; repo pushed 2026-11-07
- litellm - release train refresh: v1.108.1 and v1.108.2 stable patches, v1.107.0-rc.3 with the cosign key replaced after the Oct 21 incident, v1.110.0-dev.1 as train head
- llamafile - push refresh: decision-model support on /v1/systemone in CLI and chat modes joined a llama.cpp v0.8.0 sync; 0.10.6 still latest
- ollama - release train refresh: v0.40.15 went stable 2026-11-03 with mimo-v2.6, AutoTrust decision models, and Kimi Code CLI launch support
- experiential - volatile refresh: v0.7.233/v0.7.234 gateway-fix releases, PyPI 0.7.234 at 236 releases; the pricing page holds the 2026-11-01 rate rows against its FAQ's retained figures
- vercel-ai-gateway - rotation re-verification: all seven cited sources live with every figure held; catalog count corrected to the docs' 400+ models from 70+ providers
- chutes - rotation re-verification: PAYG-only page state, the $2.25/hour GPU rate, and every cited model rate held; the featured table grew to 18 models with GLM-5.3 newly listed
- kimi-code (model-access) - rotation re-verification: both membership ladders unchanged; the OpenRouter listing took its twenty-first move; K3.1 Pro open weights shipped to Hugging Face and the K2-series cutoff is now in effect
- kimi-code (harnesses) - mirrored the twenty-first recorded move (input leg eased to $0.192 with the output held at $16.00, the input floor at a new low of $0.1719/$9.7471 on Relace's fp4 row), appended the Price history row, refreshed the usage sentence, and corrected the K3.1 Pro caution to the open weights having shipped
- glm-coding-plan, llm-gateway, localai, magnitude: rotation re-verification or push-date refresh, nothing moved
- model-access: matrix Vercel models cell corrected to the docs' 400+ models figure
- zed: rotation re-verified, all cited URLs live, prices unchanged, llm-providers docs now route through model-access paths
- bb: rotation re-verified, desktop-v0.49.0 (Nov 6) recorded as latest stable with full written site-changelog reliability notes, and the changelog's mobile front recorded (iOS app in the update path, Android alpha since 0.45.0)
- codenomad: rotation re-verified, stable still v1.0.1 with the dev train daily through Nov 7, OpenCode recommended version moved to 2.2.1 on main, npm 881 for the month ending 2026-11-07
- trae: rotation re-verified, the $20/$60/$200 ladder and BytePlus path unchanged, trae-agent still stalled at its Feb 5 push
- codeg: recorded v0.42.1 (Nov 6) as the new latest release
- px0: recorded v0.1.21 (Nov 6) resuming the release train after the pause
- vscode-copilot: recorded 1.145.2 (Nov 6) as the current servicing patch above 1.145.1
- whiteboard: volatile re-verified, no release after v0.5.1, repository pushed 2026-11-06
- openchamber: volatile re-verified on the pushed signal, no release after v2.6.1, no note change
- surface-feature-matrix: moved the bb Mobile surface cell on the product changelog's mobile evidence; no membership change
- surfaces: OpenBitFun's Cowork-style general-workspace rejection holds on the sixth q#14 re-read; no credible surfaces-dedicated source has appeared
- surfaces: agentastic.dev reported for orchestration on the terminal-family ruling, thin record at an 11-point Show HN, resolution pending
- crowdstrike-falcon-guardian - Fal.Con Europe 2026-11-03 rework: Falcon AI Gateway generally available and included with Guardian, agent-level hooks shipped for Claude Code, OpenAI Codex, GitHub Copilot, OpenCode, and Cursor, an AI kill switch landed at gateway enforcement, and the 99 percent efficacy and sub-100-millisecond claims no longer appear on the product or solution pages
- cordum, ibm-contextforge, microsoft-agent-governance-toolkit: full rotation re-verification, every cited URL live, no drift beyond the corrected push dates and counters
- openappa - recorders line added: PR 790 (merged 2026-11-06, no release yet) replaces the runtime file ledger with recorders, so a file keeps its security label across sessions and Claude Code Bash calls run confined to declared files and hosts; breaking, preload pinned to Claude Code 2.1.266
- settlebridge - bounty marketplace now covers private and internal GitHub repositories (PR 32, 2026-11-07); commit count moved from 76 to 178 on main
- cc-safety-net - npm train moved to 2.7.4 (2026-11-06): PowerShell here-strings with CR line breaks and expansions now parse
- paperclip - release line moved to v2026.1106.0 (2026-11-06), fourteen days after v2026.1023.0; the backlog fell to 6,488
- agent-control, agentfield, hol-guard, speakeasy-gram, tracekit, vestige: volatile refreshes in place, counters and push dates moved, no release-line or content changes
- control-planes-feature-matrix: twelve volatile status cells refreshed and the CrowdStrike cells updated for the Fal.Con Europe announcements

## (daily refresh, 4)

- agent-qa - recorded the v0.3.0 (2026-11-06, remote runner and first-party plugin support), v0.3.1, and v0.3.2 releases and refreshed the activity line; npm tracks 0.3.2
- langfuse - recorded the v4.76.0 release (2026-11-05)
- reviewbench - recorded the aka-code-reviewer and zeroshot-code-review onboardings (PRs 103 and 106, both merged 2026-11-06, twenty-two total)
- verbatim-citation-gate - recorded the unreleased 2026-11-06 numeric-normalization gate fixes on main; v0.3.1 still newest, PyPI absence re-verified
- workshop - recorded the v0.1.24 release (2026-11-05, third-party dependency updates and security patches)
- frontierharness-eval, verbatim-citation-gate, workshop: rotation re-verification, every cited URL re-fetched and live
- mempabench - held report-only per the coordinator handoff, benchmark-kind candidate parked pending the owner's agent-benchmarks family ruling (q#85), 2 stars, arXiv 2610.30485, MIT
- evaluation-review: matrix maturity cells refreshed (agent-qa v0.3.2, Langfuse v4.76.0, Workshop six months); volatile push-date refreshes across agent-qa, langfuse, phoenix, reviewbench, hermes-gate, hunk, kiln, kitaru, plannotator, deepeval, harnesstax
- session-analytics: matrix column-alignment defect fixed: NovaFabric's 2026-10-28 cells had been inserted in the agents-observe position, shifting agents-observe through Memex one slot right in all eleven body rows; every row re-derived cell by cell against the member notes and re-aligned to the sorted header, verified by content fingerprint
- agentsview: volatile re-verification recorded the unreleased `agentsview why` provenance command (merged 2026-11-06), the token-usage docs flipping to schema version 6 (ninth recorded flip), forks to 680, open issues to 132
- agenttrace, claude-devtools: rotation re-verification, every cited URL live, no release movement
- coucou: recorded Windows and Linux 0.5.9 and Android 0.4.0 (2026-11-03) and the pending 0.6 wave merged 2026-11-06; forks 1,151, open issues 456
- memex: rotation re-verification recorded v0.32.0 (2026-11-02, secret redaction at index time and on egress) and v0.33.0 (2026-11-06, repeatable --session-id), gained the redaction sentence and projects-and-tags and Rex surfaces, engine count corrected to nineteen
- token-monitor: recorded v0.85.0 through v0.87.0 (2026-11-03 to 11-06); README coverage moved to 51 named tools plus a third-party APIs aggregate
- tokscale: recorded v4.21.0 (2026-11-02) and the unreleased leaderboard spoofer withholding; source table moved to 62 named clients, trailing-month downloads 131,512
- context7: released @upstash/context7-mcp@4.5.1 and ctx7@0.10.0 (both 2026-11-06: teamspace notice kept in results; Pi and T3 Code setup released), trailing-month npm 3,176,334 to 2026-11-07, pricing re-verified unchanged
- entroly: recorded main's unreleased 2026-11-07 batch (cascade-policy risk tests, STAVE fact alignment, EICV calibration, host fixes); PyPI and npm unchanged at 1.0.88
- graft: recorded main's unreleased 2026-11-06 share-session commit; npm unchanged at v0.28.1
- headroom: recorded main's unreleased 2026-11-06 to 11-07 fix batch (Vertex diagnostics, doctor IPv6 loopback, HF BatchEncoding token-count fix); no release since v0.43.1
- repowise: recorded main's unreleased 2026-11-06 merge train (sixteen community fixes plus Maven-coordinate linking); no release since v0.60.1
- code-review-graph: rotation re-verification, every cited URL intact; recorded the 2026-11-06 removal of automatic staging-to-testing promotion
- serena, vanna: rotation re-verification, every cited URL intact, no claim changed
- repomix, sol-pi, toon-format: pushed signals resolved as unreleased, side-branch, or CI-only activity, notes untouched
- context-engines: NeuroCode (Arkz-Deepak/neurocode, 0 stars, MIT) rejected on the thin-record tail, context-engine kind confirmed per the protocols handoff, self-run benchmarks only; reopen on traction or coverage
- agent-sandbox: recorded v1.0.10 (2026-11-06, router and sandboxd changes plus a NetworkPolicy rework scoping sandbox egress but opening API-server ports, breaking ProbeAddr removal), release count to thirty, pushed 2026-11-07
- brig: recorded v0.5.1 (2026-11-03, aws route preset) and v0.5.2 (2026-11-06, gcp route preset, breaking herd-to-herdr rename), pushed 2026-11-06
- cubesandbox: post-release train through 2026-11-07 (Sandbox.is_running, template QoS, cold-boot volumes, CubeOps CLI), issues to 207, pushed 2026-11-07
- drop: recorded v0.3.2 (2026-11-05, the tilde-handling security fix credited to Aditya Sai), pushed 2026-11-06
- e2b: recorded 2.63.0 (2026-11-04, strict pause with typed guest-unresponsive resume error) and the HTTP/1.1 file-transfer patches, pricing unchanged as of 2026-11-07
- fence: out-of-rotation correction, main quiet since 2026-09-10 and no release since v0.1.67; the prior reading had misread repo-level pushes as main commits; Status moved to quiet-on-main watch
- greywall, landrun, aigate, clawk: rotation re-verification, every cited URL alive, quiet windows extended, Fence compared-to claims updated for the parent's quiet train
- microsandbox: recorded v0.7.13 (2026-11-01, the Homebrew repair) and the train through 2026-11-07, pushed 2026-11-07
- nemoclaw: volatile refresh through 2026-11-07 (OpenShell 0.1.7 provenance recognition, cold startup waits), issues to 835, still exactly one prerelease
- nono: train through 2026-11-06 (plain-HTTP endpoint enforcement and network approval, NEP 0010), advisory record holds at fifteen, issues to 300
- nvx: prerelease train to 2026-11-07 (issue-fixer agent folded into a shipped skill), pushed 2026-11-07
- opensandbox: recorded release-1.2.1-rc.1 (2026-11-04, OSEP-0025 Credential Vault Exchange, atomic uploads, digest-validated snapshots), pushed 2026-11-07
- openshell: recorded v0.1.7 (2026-11-03, breaking native-OCSF-1.8.0-only telemetry) and the post-release train, issues to 629, pushed 2026-11-07
- overalls: launch moved from 3 November to 12 November 2026 per the README banner, trunk commits to 2,377, still no release or tag
- aicontainer: routine volatile refresh, downloads 517 to 476, main quiet since 2026-10-14, no Changes bullet
- sandboxing-feature-matrix: maturity cells refreshed for nine members with no membership change; audit prose parentheticals corrected; star figures untouched
- addy-osmani, harper-reed: rotation re-verification, all cited URLs live, posting claims stand
- geoffrey-huntley: rotation re-verification, all 21 cited URLs live, no new essays between 2026-10-21 and 2026-11-07, pause recorded in Status
- hamel-husain: Maven evals course repriced $4,200 -> $4,500, price history row appended, next cohort January 30 to March 27, 2027; the O'Reilly book page serves a 403 bot wall, citation kept with the wall stated
- indydevdan: Monday streak extended to 21 through the 2026-11-02 upload, view counts refreshed
- people-and-publications: matrix cadence cells moved for indydevdan and geoffrey-huntley, membership unchanged at thirty-three

## (daily refresh, 5)

- hermes: recorded v0.23.0 "The Hearth Release" (2026-11-05, about 1,950 merged pull requests since v0.22.0: gateway-served browser-tab Webapp, a single gateway owning each session, self-updating plugins, price tags and spend limits on paid MCP tools, plugin catalog past 1,050)
- nanobot: recorded v0.4.1 (2026-11-06 on GitHub and PyPI: TLS verification, sandboxed-shell and workspace-permission fixes, idle-maintenance unblocking, file-context graph)
- octop: recorded v1.0.3b3 (2026-11-04, PostgreSQL-backed skills and knowledge bases, slim Docker images) and v1.0.3b4 (2026-11-06, default Docker image flipped to slim)
- openclaw: recorded v2026.11.1 (2026-11-05, the recovery-and-reliability train) taking the latest-stable spot, with v2026.11.5-beta.1 next
- openwork: recorded v0.18.65 (2026-11-05); pricing churned a twentieth time (two add-ons returned, grandfather clause gone)
- langbot: rotation re-verification, v4.11.4 confirmed still latest, release narrative corrected to include the skipped v4.11.3 (2026-10-23)
- eigent, private-gpt: rotation re-verification, releases and pricing unchanged, counts refreshed
- assistant-runtimes: matrix re-verification moved eight status cells; comma (AFK-surf/Comma, 515 stars) rejected on the below-floor thin-record line, the OpenAgent precedent; openhuman held, q13 trigger unfired (42,578 vs 394,100 on 2026-11-07)
- automated-research-feature-matrix: corrected the Harmonic Aristotle and InternAgent columns transposed at the 2026-10-29 insertion across seven rows, found by checking one distinctive cell per column against each member's note
- harmonic-aristotle: recorded the October 27 pivot to C and Verilog code verification with CI integration and case studies, and moved pricing from free-with-an-account to consumption-based self-serve with amounts unpublished
- kapso: recorded the v0.9.0 release (November 6), its checkpoint-schema-7 break stranding 0.8.0 campaigns, the widened deploy targets, and the first-on-all-three-RelBench-leaderboards claim
- openresearch: recorded the v0.2.25 release (November 7) and its project-sharing, alphaXiv-sign-in, and chat-search train
- local-deep-research: rotation re-verification passed with all references live; recorded the November 5-6 commits, the README egress-scope documentation, and 1,385 open issues
- mole: recorded the November 6 unreleased report-corrections batch (correction-first openings, no derived arithmetic, range-aware contradiction downgrades)
- autocontext: recorded the CPU milestone wave through milestone 20
- aris: moved the push date to November 6, a documentation-only refresh
- math-inc-gauss, openai-deep-research, pion: rotation re-verification, all references re-fetched and current
- gpt-researcher and coral: updated fields aligned to their newest Changes bullets
- kodus - recorded 2.2.12 (November 5, GitHub App-linking fix), push moved to November 6, prices unchanged
- OpenCodeReview - recorded v1.13.6 (November 5: Starlark and C# review, model-provider showcase page), ledger 144 releases
- qodo - recorded the repricing: pricing page now $0.015 per credit with 2,000/4,000/16,000-credit packs (was $0.012 with 2,500/5,000/20,000), homepage FAQ still showed the old rate same-day
- roborev - push moved to November 7, v0.74.0 still latest
- OpenReview, Gito: rotation re-verification, no factual drift
- code-review: refreshed the matrix's OpenCodeReview maturity cell (144 releases) and Qodo pricing cell ($0.015 per credit); kodustech list stays unwired, staleness re-confirmed (pushed_at 2026-06-28, 54 stars)
- agent-native: npm stable crossed the 0.3.x train to 0.4.3 (2026-11-06) with pinned downloads at 6,052; the pack held at 25 skills, the intro count corrected 24 to 25
- agent-plugins: spec repo pushed 2026-11-06, main now carries a v1.1.0 Working Draft while the published site stays v1.0.0
- anthropic-agent-skills: document skills (docx, pdf, pptx, xlsx) moved to Apache-2.0 (commit 62, 2026-11-06), resolving the source-available split
- gepa: main pushed 2026-11-07 (v0.1.4 positional-order fix, DSPy ReAct trace fix), open issues 154
- opencode-skills-and-plugins: skills docs page dropped the OpenCode-1 metadata disclaimer and documents the v2 catalog surface directly; note re-grounded, conclusion unchanged
- sepia: v0.13.0 shipped 2026-11-03 (rule consolidation, three papers into the ledger, check 8 validator refusal), now the latest tag
- skillmd: v1.3.5 shipped 2026-11-03 (35th version); index 1,132,752 with reviewed holding 1,947
- skillopt: repo News announces NeurIPS 2026 Spotlight acceptance and a Microsoft Global Hackathon first place; main pushed 2026-11-07, 391 commits past v0.2.0
- skills-md: npm CLI crossed to 0.10.94 (212 versions); pricing re-verified unchanged
- skills-sh: rotation re-verification; find-skills 4.2M, hyperframes holdings down to five of the top twenty-one, skillregistry.io timing out again (third outage)
- skillspector: release train moved to v2.13.3 (2026-11-02) and v2.13.4 (2026-11-04)
- snyk-agent-scan: v0.6.11 shipped 2026-11-06; Snyk's pricing page redesigned to the credit-based enterprise model, price history row appended with the figures marked no-longer-statically-verifiable
- skills: matrix moved the Anthropic open-source cell to supported, the Snyk versioning cell to v0.6.11, the SkillMD as-of dates, and the Agent-Native npm figure to 0.4.3
- agent-skills-open-standard: rotation re-verification, every cited URL re-fetched live, claims held a fourth time, the 79-agent count re-dated

## (daily refresh, 6)

- ai-dlc: recorded the stable v2.11.3 release (2026-11-04, the native aidlc binary rebuilt on Bun 1.4.4 fixing a CVE, team Units end to end)
- ai-unified-process: recorded the fourth release wave (2026-11-04: aiup-core v3.13.0 with a spec-review warning, Vaadin v3.5.0)
- arc-kit: recorded the v6.19.0 release (2026-11-02: uncertainty envelopes on chart marks, the SBOM command rebuilt on SLSA v1.2)
- archcore: corrected the implementation language back to Go (the v0.15.3 release notes, README badge, and dev-branch tree, against the default branch the 10-27 TypeScript correction had read) and recorded the v0.15.0 through v0.15.3 release train; out-of-rotation re-verification
- gsd: recorded the successor's v1.20.0 and v1.20.1 releases (2026-11-03 and 05) and refreshed the npm figures
- mycelium: recorded the v0.333.0 through v0.334.1 releases (all 2026-11-06, the eighth xai-check and the first blind-prediction faithfulness audit)
- openspec: recorded the v1.16.1 release (2026-11-07: unrecognized delta sections now fail validation) and refreshed the download count (3,986,826)
- pdd-kit: recorded the v0.39.0 and v0.40.0 releases (2026-11-02 and 06) and refreshed the npm figure
- spec-kit: recorded the v1.2.6 through v1.3.2 releases (2026-11-02 to 06: Quality Gates v0.6.0, the Assure extension joining the catalog)
- superpowers: rotation re-verification (all cited sources alive) plus v7.0.2 (2026-11-03) recorded
- tessl, vibe-coding-prompt-template: rotation re-verification, all cited sources alive, no facts changed beyond the npm CLI figure
- spec-driven-development: matrix adoption row as-of dates moved and the Archcore Kind cell corrected back to a Go CLI; category index one-liners corrected
- beads: recorded v1.3.3 (2026-11-03, security and data-loss fixes) and v1.3.4 (2026-11-04, the GT_ROOT restore); counts refreshed
- agentrq: recorded the v0.15.3 to v0.15.7 train (2026-11-02 to 2026-11-05) and refreshed counts
- ccpm: recorded the 2026-11-04 activity resumption on main (the #1031 sync fix and the README Neon guide); still no tagged releases
- its-a-plan: recorded v1.10.0 (2026-11-05: agent workflow canvases, project templates and public gallery, run queues, system roles); counts refreshed
- kata: recorded v0.21.0 (Docbank attachments, document-version evidence closes, kata move) and v0.22.0 (TUI deletion with undo, cron editing, 100k-issue scale fixes); counts refreshed
- operon: recorded 4.5.0 (Korean, Status Dock sections, session history, kanban task pool); counts refreshed
- ordewell: recorded v0.12.4 to v0.13.0 (the Windows line, Cursor as a fourth runner over ACP, source-only desktop app announcement) and corrected the runner count
- backlog-md, epiq, pad: counts-only refreshes, releases unchanged
- task-master, veritas-kanban: rotation re-verification with every cited URL re-fetched, all claims hold
- task-management: matrix release and activity cells refreshed; ralph-tui (2,461 stars) rejected on the Ralph-loop wrapper ruling, orchestration-kind; Nimbalyst (2,036 stars) report-only to cross-category; q24 evidence refreshed, the TomzxCode/agentic rejection stands
- software-factory: added Open SWE and Open-Inspect (12 members); matrix extended to twelve columns, index updated
- open-swe: created from the 2026-11-06 routed admission; Robin rebrand, Managed Deep Agents requirement, installation-wide credential trust model, 111-point launch thread recorded, 11,177-star baseline
- open-inspect: created from the varun1505 2026-11-05 sweep; per-session scoped sandbox credentials, no release train, zero independent coverage recorded, 3,467-star baseline
- zeroshot: recorded the v11.0.0 major (delivery evidence separated from graph output) and v11.1.0; README rewrite repositions as a multi-agent orchestration framework
- loki-mode: train to v11.138.0 with the acceptance-shift and never-stop changes recorded, npm in sync, Docker pulls 324,934
- har: recorded the v1.16.1 release (2026-11-03, Playwright 1.64.1 pin); rotation re-verified, pricing unchanged
- fusion: rotation re-verification, npm latest still 0.77.0 with the stable channel nine-plus weeks behind, downloads 1,843 per month
- fabro, machinist, ouroboros, superplane: volatile refreshes, no bullets (routine volatile); SuperPlane's largest-member claim dropped now that Robin leads the category
- software-factory: the varun1505 2026-11-02 and 2026-11-05 sweeps resolved; DevBox held below the bar on the Orbi coverage precedent, Helix rejected on the no-standard-license line, DevIntern on the thin tail, Foundry and Coasts dormant under the CodeMachine precedent, Volt and Bottleneck second listings carried
- benchmark-radar: v0.14.0 stable release published 2026-11-03 merged the two sites into one; discovery widened to 41 public sources; badge refreshed to 45,809 records over 103 days as of 2026-11-07
- bestvaluemodel: November diff-feed events recorded (the 11-03 mass re-score, eight new rows, the 11-07 Nemotron price cut); repository still unlicensed
- epoch-ai: rotation re-verification with every cited URL re-fetched live: team 62, AI-models explorer over 3,700 models, stamps through Nov 7
- lmarena: rotation re-verification with every cited URL re-fetched live: two 30 October posts recorded, redirects and arena-rank dormancy unchanged
- trackers-and-leaderboards: matrix cells refreshed and alignment verified (one distinctive cell per member column); the dedicated-source check re-ran, the category still keeps none, fifteen candidates rejected on standing lines
- cross-category: agentastic.dev (Agentastic.dev mac multi-agent IDE, reported by the surfaces worker on the terminal-family ruling) rejected on the thin-record tail: closed product with no public code repository (the org holds a 1-star homebrew tap and a 0-star quickstart), an 11-point January Show HN, zero coverage since
- cross-category: the task-management worker flagged three main-corpus files outside agents/ carrying broken internal links, queued for the owner
- no self-directed essay this run: every candidate topic read as covered by the existing essays and trackers, so the run closed maintenance-only beyond the recorded admissions and fixes