[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BGJWV0QAEF6E4M1DDFDP7F":3,"skill-same:01M7BGJWV0QAEF6E4M1DDFDP7F":49},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":29,"commitSha":30,"skillDirectory":31,"forksCount":14,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":34,"install":35,"resources":48},"01M7BGJWV0QAEF6E4M1DDFDP7F","ai-release-hunter",null,"Daily digest of AI releases from Anthropic, Meta, Vercel Labs, OpenAI and Google GitHub repos, the Claude Cowork changelog, the claude.dev blog and Hacker News, deduplicated against a local state file and sent by email. Use when the user asks for a daily AI news digest, to scan AI labs for new releases, or to set up a scheduled release-hunter job.","origin-labs-ai\u002Fcode","origin-labs-ai","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F347654492?v=4","01M7B0MAD1JMVZ6ME9891AYYYG","Origin","01M7B0MADP3AZS42R4NEF1JTS8",0,"MIT","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fpayloads\u002Fskills\u002Fai-release-hunter\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: ai-release-hunter\ndescription: Daily digest of AI releases from Anthropic, Meta, Vercel Labs, OpenAI and Google GitHub repos, the Claude Cowork changelog, the claude.dev blog and Hacker News, deduplicated against a local state file and sent by email. Use when the user asks for a daily AI news digest, to scan AI labs for new releases, or to set up a scheduled release-hunter job.\nlicense: MIT\nallowed-tools: Bash Read Write WebFetch\nmetadata:\n  author: Daniel Avila\n  version: \"1.0.0\"\n  tags: \"ai-news, releases, github, hacker-news, digest, email, anthropic, openai, google, meta, vercel\"\n---\n\n# AI Release Hunter (Anthropic + Meta + Vercel + OpenAI + Google + Hacker News)\n\nScans the public AI repos of Anthropic, Meta, Vercel Labs, OpenAI and Google on GitHub, plus the Claude Cowork changelog, the claude.dev blog and Hacker News, and emails a digest of the day's relevant movements. Designed to run once a day as a scheduled job (cron, a Claude Code routine, or `\u002Floop`).\n\n## Configuration\n\nSet these before the first run. Every path is relative to the working directory unless you choose otherwise.\n\n| Setting | Default | Purpose |\n|---------|---------|---------|\n| `RECIPIENT_EMAIL` | *(required)* | Address that receives the digest |\n| `SEND_COMMAND` | *(required)* | Command that sends an email given a recipient, subject and body file (e.g. a Gmail CLI, `mail` or `sendmail`; to send through an MCP tool instead, add it to `allowed-tools`) |\n| `STATE_DIR` | `.release-hunter\u002Fstate\u002F` | Deduplication state (`seen.json`) |\n| `SNAPSHOT_DIR` | `.release-hunter\u002Fsnapshots\u002F` | Page snapshots (Vercel Labs, Cowork changelog) |\n| `OUTPUT_DIR` | `.release-hunter\u002Foutput\u002F` | Email bodies and run logs |\n| `EMAIL_LANGUAGE` | Spanish | Language of the email |\n| `EMAIL_SUBJECT` | `IA: novedades del día (Anthropic + Meta + Vercel + OpenAI + Google + Hacker News)` | Subject line |\n| `AUDIENCE` | the user's audience | Who the \"why it matters\" lines are written for (e.g. a YouTube channel, a team) |\n\nSuggested schedule: daily at ~08:00 in the user's time zone.\n\nAdd `.release-hunter\u002F` to `.gitignore` so state, snapshots and email bodies are never committed. The scan makes well over 60 GitHub API calls per run, so a `GITHUB_TOKEN` is strongly recommended (see below).\n\n## Watched sources\n\n### Anthropic (GitHub org `anthropics`)\n- anthropics\u002Fclaude-code\n- anthropics\u002Fclaude-cookbooks\n- anthropics\u002Fanthropic-sdk-python\n- anthropics\u002Fanthropic-sdk-typescript\n- anthropics\u002Fcourses\n\nAlso list the org's repos (`https:\u002F\u002Fapi.github.com\u002Forgs\u002Fanthropics\u002Frepos?per_page=100`) and consider any other with recent relevant activity (agent SDKs, examples, tooling).\n\n### claude.dev blog (https:\u002F\u002Fclaude.dev\u002F)\nAnthropic's technical blog with tips, techniques and POVs from its developers for building with Claude. Fetch it with WebFetch and take the list of recent posts (date + title + link). Compare against the already-reported ones in `STATE_DIR\u002Fseen.json` under the key `claude_dev_posts` (list of slugs or URLs; create it if missing). Report each new post as a highlighted item: title, date, one line on why it matters to `AUDIENCE` + link to the post. Add the new ones at the end.\n\n### Meta AI (GitHub orgs `meta-llama` and `facebookresearch`)\n- meta-llama\u002Fllama-models\n- meta-llama\u002Fllama-stack\n- meta-llama\u002Fllama-agentic-system\n- meta-llama\u002Fllama-toolchain\n- meta-llama\u002Fllama-cookbook\n- meta-llama\u002Fpurplellama\n- meta-llama\u002Fcodellama\n- facebookresearch\u002Fsegment-anything\n- facebookresearch\u002Fsegment-anything-2\n- facebookresearch\u002Fdinov2\n- facebookresearch\u002Fdinov3\n- facebookresearch\u002Ffaiss\n- facebookresearch\u002Faudiocraft\n- facebookresearch\u002Fspiritlm\n- facebookresearch\u002Fvggt\n- facebookresearch\u002Fvjepa2\n- facebookresearch\u002Fxformers\n\nIf any repo returns 404, skip it and continue. Also list both orgs' repos and add up to ~15 more with a push in the last 30 days whose name or description fits AI: new models, agents, MCP, vision, audio, multimodal, diffusion, segmentation, embeddings, fine-tuning, RL. Ignore internal infra repos, consumer product demos or archived ones.\n\n### Vercel Labs (org `vercel-labs`; experiments at https:\u002F\u002Fvercel.com\u002Flabs)\n- vercel-labs\u002Fagent-browser (browser automation CLI for AI agents)\n- vercel-labs\u002Fdeepsec (security harness for coding agents)\n- vercel-labs\u002Fjson-render (Generative UI framework)\n- vercel-labs\u002Fskills (open agent skills ecosystem CLI)\n- vercel-labs\u002Fportless\n- vercel-labs\u002Ffx (unix-like coding agent)\n- vercel-labs\u002Fvgpu\n- vercel-labs\u002Fscriptc, vercel-labs\u002Fwterm, vercel-labs\u002Femulate, vercel-labs\u002Fnative, vercel-labs\u002Fcallscript, vercel-labs\u002Frun, vercel-labs\u002Fvisual-json, vercel-labs\u002Fzerolang, vercel-labs\u002Fphase, vercel-labs\u002Fwebreel, vercel-labs\u002Fopensrc\n\nIf any repo returns 404, skip it and continue. Also list the org's repos and add those with a push in the last 30 days not already in the list.\n\nLabs page updates: fetch https:\u002F\u002Fvercel.com\u002Flabs and compare the experiment list (name + category: Labs products \u002F Active experiments \u002F Past experiments) against the snapshot at `SNAPSHOT_DIR\u002Fvercel-labs-experiments.json`. Report any new experiment or category change as a highlighted item. Update the snapshot at the end.\n\n### OpenAI (GitHub org `openai`)\n- openai\u002Fcodex (coding agent CLI)\n- openai\u002Fcodex-security (Codex Security CLI + TypeScript SDK)\n- openai\u002Fcodex-action (GitHub Action for Codex)\n- openai\u002Fopenai-agents-python (agent framework, Python)\n- openai\u002Fopenai-agents-js (agent framework, JS)\n- openai\u002Fopenai-guardrails-python\n- openai\u002Fopenai-guardrails-js\n- openai\u002Ftunnel-client (Secure MCP Tunnel client)\n- openai\u002Fopenai-cua-sample-app (computer use via API)\n- openai\u002Fsymphony (autonomous implementation runs)\n- openai\u002Fopenai-cli (official CLI)\n- openai\u002Fplugins, openai\u002Fcommunity-plugins\n- openai\u002Ffence\n\nIf any repo returns 404, skip it and continue. Also list the org's repos and add up to ~10 more with a push in the last 30 days whose name or description fits AI: agents, codex, MCP, skills, computer use, guardrails, plugins, sandboxing. Ignore internal infra repos, consumer product demos or archived ones.\n\n### Google (GitHub org `google`)\n- google\u002Fax (Google's open agentic orchestration runtime; priority: releases, roadmap, API changes)\n- google\u002Fadk-python (Agent Development Kit, Python)\n- google\u002Fadk-js (Agent Development Kit, JS)\n- google\u002Fgemini-cli (Gemini coding agent CLI)\n- google\u002Fa2a (Agent2Agent protocol)\n- google\u002Fgenai-python (Gemini SDK for Python)\n- google\u002Fgenai-js (Gemini SDK for JS)\n\nIf any repo returns 404, skip it and continue. Also list the org's repos and add up to ~10 more with a push in the last 30 days whose name or description fits agentic AI: agents, orchestration, A2A\u002FMCP, Gemini, evals, tool use. Ignore internal infra repos, consumer product demos, Android, Flutter or archived ones.\n\nGoogle items are reported as-is, **with no Claude\u002FAnthropic comparison angle**.\n\n### Claude Cowork changelog (https:\u002F\u002Fclaude.com\u002Fdocs\u002Fcowork\u002Fchangelog)\nThe page is public and organizes changes by date (`YYYY-MM-DD` headings). Compare the date sections against the snapshot at `SNAPSHOT_DIR\u002Fcowork-changelog-snapshot.json` (format: date → `{\"sha256\": hash of that section's content}`).\n\nExact hash method: download the HTML with curl, strip `\u003Cscript>`\u002F`\u003Cstyle>` tags, convert remaining tags to newlines, drop empty lines, and split on lines that are only a date `YYYY-MM-DD`. The sha256 is computed over each section's text, with no HTML-entity unescaping. Keep the method identical across runs or every section will look changed.\n\nReport only new dates or dates whose content changed since the snapshot, with the most relevant bullets (features, behavior changes, new Cowork\u002FCode\u002F3P settings; skip minor bug fixes unless important) + the changelog page link. Update the snapshot at the end.\n\n### Hacker News (https:\u002F\u002Fnews.ycombinator.com)\nFind stories from the last ~24 hours relevant to Anthropic, Meta AI, Vercel, OpenAI, Google AI, Claude\u002FClaude Code, coding agents and MCP. Use the public Algolia API (no auth), with parallel requests:\n- `GET https:\u002F\u002Fhn.algolia.com\u002Fapi\u002Fv1\u002Fsearch?query=\u003Cterm>&tags=story&numericFilters=created_at_i>\u003Ctimestamp_24h_ago>,points>40` for each keyword: `Claude Code`, `Codex`, `Cursor`, `MCP`, `Anthropic`, `OpenAI`, `Google`, `Vercel`, `Llama`, `AI agent`\n- Also check the current front page: `GET https:\u002F\u002Fhn.algolia.com\u002Fapi\u002Fv1\u002Fsearch?tags=front_page` and keep the relevant stories.\n\nFilter noise: only stories with real traction (meaningful points and comments) and direct relevance to `AUDIENCE` (launches, technical debates on coding agents, model comparisons, new tools). Deduplicate by storing already-reported story IDs in `STATE_DIR\u002Fseen.json` under the key `hn_story_ids`. Only report unseen stories; add the new IDs at the end.\n\n## What to look for in each repo\n\nUse the public GitHub API, no auth, with parallel requests (e.g. `xargs -P16`). Unauthenticated calls are limited to 60 requests\u002Fhour per IP; if a `GITHUB_TOKEN` is available in the environment, send it as a bearer token to raise the limit, and never write it to any output file.\n\n1. **New releases:** `GET \u002Frepos\u002F{owner}\u002F{repo}\u002Freleases`; take those published since the last check (filter by publication date, not just tag equality).\n2. **Relevant commits** on the main branch since the last check: `GET \u002Frepos\u002F{owner}\u002F{repo}\u002Fcommits?since=\u003CISO-8601>`. Filter noise (dependabot, CI, typo fixes); keep features, behavior changes, new commands\u002Fflags, deprecations, new models and important fixes.\n3. **Notable changes** in CHANGELOG or README when applicable.\n\n## Deduplication\n\nStore in `STATE_DIR\u002Fseen.json` (create if missing) the last release tag and last reviewed commit SHA per repo, keyed by `owner\u002Frepo`. Only report what's new since the last mark; update the mark at the end. On the first run there is no mark: report only the last 24 hours instead of the full history.\n\n## Email format\n\n- Written in `EMAIL_LANGUAGE`, direct and no filler.\n- Grouped by company and repo, each item with one line on why it matters to `AUDIENCE` + link to the release or commit.\n- When a Meta, Vercel or OpenAI item has a Claude\u002FAnthropic comparison angle (e.g. Codex vs Claude Code, Llama vs Claude, agent skills, MCP), mention it: contrasts are useful to the reader.\n- A `Hacker News` section: per story, title, points\u002Fcomments, one line on why it matters + link to the discussion (`https:\u002F\u002Fnews.ycombinator.com\u002Fitem?id=\u003Cid>`).\n- If nothing relevant: a one-line email saying there were no relevant movements today.\n\n## Sending\n\nWrite the body to a durable file first, `OUTPUT_DIR\u002Femail-body-YYYY-MM-DD.txt`, never to `\u002Ftmp` (it can be wiped between turns of a scheduled job). Then send it with `SEND_COMMAND`, for example:\n\n```bash\n\u003Csend-command> --to \"$RECIPIENT_EMAIL\" --subject \"$EMAIL_SUBJECT\" --body \"$(cat .release-hunter\u002Foutput\u002Femail-body-YYYY-MM-DD.txt)\"\n```\n\nThe user must authorize the recurring send when setting up the job; do not send to any address other than `RECIPIENT_EMAIL`. If sending fails, record it in the run log.\n\nIf the GitHub API fails or returns a rate limit, wait a minute and retry once; if it still fails, send the email noting which sources couldn't be checked today.\n\n## Wrap-up\n\nAppend a run summary (sources checked, items reported, failures) to `OUTPUT_DIR\u002Fruns\u002FYYYY-MM-DD.md`, and save the updated `seen.json` and snapshots only after the email is sent, so a failed send is retried with the same items next run.\n","# AI Release Hunter (Anthropic + Meta + Vercel + OpenAI + Google + Hacker News)\n\nScans the public AI repos of Anthropic, Meta, Vercel Labs, OpenAI and Google on GitHub, plus the Claude Cowork changelog, the claude.dev blog and Hacker News, and emails a digest of the day's relevant movements. Designed to run once a day as a scheduled job (cron, a Claude Code routine, or `\u002Floop`).\n\n## Configuration\n\nSet these before the first run. Every path is relative to the working directory unless you choose otherwise.\n\n| Setting | Default | Purpose |\n|---------|---------|---------|\n| `RECIPIENT_EMAIL` | *(required)* | Address that receives the digest |\n| `SEND_COMMAND` | *(required)* | Command that sends an email given a recipient, subject and body file (e.g. a Gmail CLI, `mail` or `sendmail`; to send through an MCP tool instead, add it to `allowed-tools`) |\n| `STATE_DIR` | `.release-hunter\u002Fstate\u002F` | Deduplication state (`seen.json`) |\n| `SNAPSHOT_DIR` | `.release-hunter\u002Fsnapshots\u002F` | Page snapshots (Vercel Labs, Cowork changelog) |\n| `OUTPUT_DIR` | `.release-hunter\u002Foutput\u002F` | Email bodies and run logs |\n| `EMAIL_LANGUAGE` | Spanish | Language of the email |\n| `EMAIL_SUBJECT` | `IA: novedades del día (Anthropic + Meta + Vercel + OpenAI + Google + Hacker News)` | Subject line |\n| `AUDIENCE` | the user's audience | Who the \"why it matters\" lines are written for (e.g. a YouTube channel, a team) |\n\nSuggested schedule: daily at ~08:00 in the user's time zone.\n\nAdd `.release-hunter\u002F` to `.gitignore` so state, snapshots and email bodies are never committed. The scan makes well over 60 GitHub API calls per run, so a `GITHUB_TOKEN` is strongly recommended (see below).\n\n## Watched sources\n\n### Anthropic (GitHub org `anthropics`)\n- anthropics\u002Fclaude-code\n- anthropics\u002Fclaude-cookbooks\n- anthropics\u002Fanthropic-sdk-python\n- anthropics\u002Fanthropic-sdk-typescript\n- anthropics\u002Fcourses\n\nAlso list the org's repos (`https:\u002F\u002Fapi.github.com\u002Forgs\u002Fanthropics\u002Frepos?per_page=100`) and consider any other with recent relevant activity (agent SDKs, examples, tooling).\n\n### claude.dev blog (https:\u002F\u002Fclaude.dev\u002F)\nAnthropic's technical blog with tips, techniques and POVs from its developers for building with Claude. Fetch it with WebFetch and take the list of recent posts (date + title + link). Compare against the already-reported ones in `STATE_DIR\u002Fseen.json` under the key `claude_dev_posts` (list of slugs or URLs; create it if missing). Report each new post as a highlighted item: title, date, one line on why it matters to `AUDIENCE` + link to the post. Add the new ones at the end.\n\n### Meta AI (GitHub orgs `meta-llama` and `facebookresearch`)\n- meta-llama\u002Fllama-models\n- meta-llama\u002Fllama-stack\n- meta-llama\u002Fllama-agentic-system\n- meta-llama\u002Fllama-toolchain\n- meta-llama\u002Fllama-cookbook\n- meta-llama\u002Fpurplellama\n- meta-llama\u002Fcodellama\n- facebookresearch\u002Fsegment-anything\n- facebookresearch\u002Fsegment-anything-2\n- facebookresearch\u002Fdinov2\n- facebookresearch\u002Fdinov3\n- facebookresearch\u002Ffaiss\n- facebookresearch\u002Faudiocraft\n- facebookresearch\u002Fspiritlm\n- facebookresearch\u002Fvggt\n- facebookresearch\u002Fvjepa2\n- facebookresearch\u002Fxformers\n\nIf any repo returns 404, skip it and continue. Also list both orgs' repos and add up to ~15 more with a push in the last 30 days whose name or description fits AI: new models, agents, MCP, vision, audio, multimodal, diffusion, segmentation, embeddings, fine-tuning, RL. Ignore internal infra repos, consumer product demos or archived ones.\n\n### Vercel Labs (org `vercel-labs`; experiments at https:\u002F\u002Fvercel.com\u002Flabs)\n- vercel-labs\u002Fagent-browser (browser automation CLI for AI agents)\n- vercel-labs\u002Fdeepsec (security harness for coding agents)\n- vercel-labs\u002Fjson-render (Generative UI framework)\n- vercel-labs\u002Fskills (open agent skills ecosystem CLI)\n- vercel-labs\u002Fportless\n- vercel-labs\u002Ffx (unix-like coding agent)\n- vercel-labs\u002Fvgpu\n- vercel-labs\u002Fscriptc, vercel-labs\u002Fwterm, vercel-labs\u002Femulate, vercel-labs\u002Fnative, vercel-labs\u002Fcallscript, vercel-labs\u002Frun, vercel-labs\u002Fvisual-json, vercel-labs\u002Fzerolang, vercel-labs\u002Fphase, vercel-labs\u002Fwebreel, vercel-labs\u002Fopensrc\n\nIf any repo returns 404, skip it and continue. Also list the org's repos and add those with a push in the last 30 days not already in the list.\n\nLabs page updates: fetch https:\u002F\u002Fvercel.com\u002Flabs and compare the experiment list (name + category: Labs products \u002F Active experiments \u002F Past experiments) against the snapshot at `SNAPSHOT_DIR\u002Fvercel-labs-experiments.json`. Report any new experiment or category change as a highlighted item. Update the snapshot at the end.\n\n### OpenAI (GitHub org `openai`)\n- openai\u002Fcodex (coding agent CLI)\n- openai\u002Fcodex-security (Codex Security CLI + TypeScript SDK)\n- openai\u002Fcodex-action (GitHub Action for Codex)\n- openai\u002Fopenai-agents-python (agent framework, Python)\n- openai\u002Fopenai-agents-js (agent framework, JS)\n- openai\u002Fopenai-guardrails-python\n- openai\u002Fopenai-guardrails-js\n- openai\u002Ftunnel-client (Secure MCP Tunnel client)\n- openai\u002Fopenai-cua-sample-app (computer use via API)\n- openai\u002Fsymphony (autonomous implementation runs)\n- openai\u002Fopenai-cli (official CLI)\n- openai\u002Fplugins, openai\u002Fcommunity-plugins\n- openai\u002Ffence\n\nIf any repo returns 404, skip it and continue. Also list the org's repos and add up to ~10 more with a push in the last 30 days whose name or description fits AI: agents, codex, MCP, skills, computer use, guardrails, plugins, sandboxing. Ignore internal infra repos, consumer product demos or archived ones.\n\n### Google (GitHub org `google`)\n- google\u002Fax (Google's open agentic orchestration runtime; priority: releases, roadmap, API changes)\n- google\u002Fadk-python (Agent Development Kit, Python)\n- google\u002Fadk-js (Agent Development Kit, JS)\n- google\u002Fgemini-cli (Gemini coding agent CLI)\n- google\u002Fa2a (Agent2Agent protocol)\n- google\u002Fgenai-python (Gemini SDK for Python)\n- google\u002Fgenai-js (Gemini SDK for JS)\n\nIf any repo returns 404, skip it and continue. Also list the org's repos and add up to ~10 more with a push in the last 30 days whose name or description fits agentic AI: agents, orchestration, A2A\u002FMCP, Gemini, evals, tool use. Ignore internal infra repos, consumer product demos, Android, Flutter or archived ones.\n\nGoogle items are reported as-is, **with no Claude\u002FAnthropic comparison angle**.\n\n### Claude Cowork changelog (https:\u002F\u002Fclaude.com\u002Fdocs\u002Fcowork\u002Fchangelog)\nThe page is public and organizes changes by date (`YYYY-MM-DD` headings). Compare the date sections against the snapshot at `SNAPSHOT_DIR\u002Fcowork-changelog-snapshot.json` (format: date → `{\"sha256\": hash of that section's content}`).\n\nExact hash method: download the HTML with curl, strip `\u003Cscript>`\u002F`\u003Cstyle>` tags, convert remaining tags to newlines, drop empty lines, and split on lines that are only a date `YYYY-MM-DD`. The sha256 is computed over each section's text, with no HTML-entity unescaping. Keep the method identical across runs or every section will look changed.\n\nReport only new dates or dates whose content changed since the snapshot, with the most relevant bullets (features, behavior changes, new Cowork\u002FCode\u002F3P settings; skip minor bug fixes unless important) + the changelog page link. Update the snapshot at the end.\n\n### Hacker News (https:\u002F\u002Fnews.ycombinator.com)\nFind stories from the last ~24 hours relevant to Anthropic, Meta AI, Vercel, OpenAI, Google AI, Claude\u002FClaude Code, coding agents and MCP. Use the public Algolia API (no auth), with parallel requests:\n- `GET https:\u002F\u002Fhn.algolia.com\u002Fapi\u002Fv1\u002Fsearch?query=\u003Cterm>&tags=story&numericFilters=created_at_i>\u003Ctimestamp_24h_ago>,points>40` for each keyword: `Claude Code`, `Codex`, `Cursor`, `MCP`, `Anthropic`, `OpenAI`, `Google`, `Vercel`, `Llama`, `AI agent`\n- Also check the current front page: `GET https:\u002F\u002Fhn.algolia.com\u002Fapi\u002Fv1\u002Fsearch?tags=front_page` and keep the relevant stories.\n\nFilter noise: only stories with real traction (meaningful points and comments) and direct relevance to `AUDIENCE` (launches, technical debates on coding agents, model comparisons, new tools). Deduplicate by storing already-reported story IDs in `STATE_DIR\u002Fseen.json` under the key `hn_story_ids`. Only report unseen stories; add the new IDs at the end.\n\n## What to look for in each repo\n\nUse the public GitHub API, no auth, with parallel requests (e.g. `xargs -P16`). Unauthenticated calls are limited to 60 requests\u002Fhour per IP; if a `GITHUB_TOKEN` is available in the environment, send it as a bearer token to raise the limit, and never write it to any output file.\n\n1. **New releases:** `GET \u002Frepos\u002F{owner}\u002F{repo}\u002Freleases`; take those published since the last check (filter by publication date, not just tag equality).\n2. **Relevant commits** on the main branch since the last check: `GET \u002Frepos\u002F{owner}\u002F{repo}\u002Fcommits?since=\u003CISO-8601>`. Filter noise (dependabot, CI, typo fixes); keep features, behavior changes, new commands\u002Fflags, deprecations, new models and important fixes.\n3. **Notable changes** in CHANGELOG or README when applicable.\n\n## Deduplication\n\nStore in `STATE_DIR\u002Fseen.json` (create if missing) the last release tag and last reviewed commit SHA per repo, keyed by `owner\u002Frepo`. Only report what's new since the last mark; update the mark at the end. On the first run there is no mark: report only the last 24 hours instead of the full history.\n\n## Email format\n\n- Written in `EMAIL_LANGUAGE`, direct and no filler.\n- Grouped by company and repo, each item with one line on why it matters to `AUDIENCE` + link to the release or commit.\n- When a Meta, Vercel or OpenAI item has a Claude\u002FAnthropic comparison angle (e.g. Codex vs Claude Code, Llama vs Claude, agent skills, MCP), mention it: contrasts are useful to the reader.\n- A `Hacker News` section: per story, title, points\u002Fcomments, one line on why it matters + link to the discussion (`https:\u002F\u002Fnews.ycombinator.com\u002Fitem?id=\u003Cid>`).\n- If nothing relevant: a one-line email saying there were no relevant movements today.\n\n## Sending\n\nWrite the body to a durable file first, `OUTPUT_DIR\u002Femail-body-YYYY-MM-DD.txt`, never to `\u002Ftmp` (it can be wiped between turns of a scheduled job). Then send it with `SEND_COMMAND`, for example:\n\n```bash\n\u003Csend-command> --to \"$RECIPIENT_EMAIL\" --subject \"$EMAIL_SUBJECT\" --body \"$(cat .release-hunter\u002Foutput\u002Femail-body-YYYY-MM-DD.txt)\"\n```\n\nThe user must authorize the recurring send when setting up the job; do not send to any address other than `RECIPIENT_EMAIL`. If sending fails, record it in the run log.\n\nIf the GitHub API fails or returns a rate limit, wait a minute and retry once; if it still fails, send the email noting which sources couldn't be checked today.\n\n## Wrap-up\n\nAppend a run summary (sources checked, items reported, failures) to `OUTPUT_DIR\u002Fruns\u002FYYYY-MM-DD.md`, and save the updated `seen.json` and snapshots only after the email is sent, so a failed send is retried with the same items next run.\n",{"name":5,"license":15,"metadata":24,"description":7,"allowed-tools":28},{"tags":25,"author":26,"version":27},"ai-news, releases, github, hacker-news, digest, email, anthropic, openai, google, meta, vercel","Daniel Avila","1.0.0","Bash Read Write WebFetch","main","1e72c52e0f3ffd72fbc6ba994fda0f4a6d63911a","payloads\u002Fskills\u002Fai-release-hunter","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai","2026-11-14 20:35:25",{"source":8,"skillName":5,"commands":36,"prompt":46,"zipAvailable":47},[37,40,43],{"tool":38,"command":39},"npx","npx skills add origin-labs-ai\u002Fcode --skill ai-release-hunter",{"tool":41,"command":42},"bunx","bunx skills add origin-labs-ai\u002Fcode --skill ai-release-hunter",{"tool":44,"command":45},"pnpm","pnpm dlx skills add origin-labs-ai\u002Fcode --skill ai-release-hunter","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fpayloads\u002Fskills\u002Fai-release-hunter\u002FSKILL.md",true,[],[50,58,66,74,82,91],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":54,"sourceHtmlUrl":55,"categories":56,"keywords":57,"publishedAt":19},"01M7BP6Z7NBQXMB8Y5ZRR7PGCF","web-search","Implement web search capabilities using the z-ai-web-dev-sdk. Use this skill when the user needs to search for real-time information from the web, retrieve up-to-date content beyond the knowledge cutoff, or find the latest news and data. Returns structured search results with URLs, snippets, and metadata.","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fweb-search\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fweb-search\u002FSKILL.md",[],[],{"skillNo":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":62,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M7BP6T13CWYHZSHGYG0RT885","web-reader","Implement web page content extraction capabilities using the z-ai-web-dev-sdk. Use this skill when the user needs to scrape web pages, extract article content, retrieve page metadata, or build applications that process web content. Supports automatic content extraction with title, HTML, and publication time retrieval.","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fweb-reader\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fweb-reader\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":70,"sourceHtmlUrl":71,"categories":72,"keywords":73,"publishedAt":19},"01M7BP6GM9Z301PPT4ZRV049AN","video-understand","Implement specialized video understanding capabilities using the z-ai-web-dev-sdk. Use this skill when the user needs to analyze video content, understand motion and temporal sequences, extract information from video frames, describe video scenes, or perform video-based AI analysis. Optimized for MP4, AVI, MOV, and other common video formats.","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fvideo-understand\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fvideo-understand\u002FSKILL.md",[],[],{"skillNo":75,"name":76,"nameZh":6,"description":77,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":78,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M7BP6ANRHQ1K7PVS11RBZCC7","Video Generation","Implement AI-powered video generation capabilities using the z-ai-web-dev-sdk. Use this skill when the user needs to generate videos from text prompts or images, create video content programmatically, or build applications that produce video outputs. Supports asynchronous task management with status polling and result retrieval.","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fvideo-generation\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fvideo-generation\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":86,"licenseFileUrl":87,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M7BP56W5SX2MMJWKMJVC66F0","skill-creator","Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.","Apache-2.0","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fskill-creator\u002FLICENSE.txt","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fskill-creator\u002FSKILL.md",[],[],{"skillNo":92,"name":93,"nameZh":6,"description":94,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":86,"licenseFileUrl":6,"sourceHtmlUrl":95,"categories":96,"keywords":97,"publishedAt":19},"01M7BP503HPG0DBCN4K0JKCA88","seo-content-writer","Use when the user asks to \"write SEO content\", \"create a blog post\", \"write an article\", \"content writing\", \"draft optimized content\", \"write me an article\", \"create a blog post about\", \"help me write SEO content\", or \"draft content for\". Creates high-quality, SEO-optimized content that ranks in search engines. Applies on-page SEO best practices, keyword optimization, and content structure for maximum visibility and engagement. For AI citation optimization, see geo-content-optimizer. For updating existing content, see content-refresher.","https:\u002F\u002Fgithub.com\u002Forigin-labs-ai\u002Fcode\u002Fblob\u002Fmain\u002Fwindows\u002Fsonar\u002Fmerger\u002Fcode-fusion-dead-20261101\u002Fskills\u002Fseo-content-writer\u002FSKILL.md",[],[]]