[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6CZ592A5T2RP7CVHNFWK6HH":3,"skill-same:01M6CZ592A5T2RP7CVHNFWK6HH":87},{"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":31,"publishedAt":52,"linkOnly":53,"markdownContent":54,"markdownBody":55,"frontMatter":56,"refName":64,"commitSha":65,"skillDirectory":66,"forksCount":67,"repositoryHtmlUrl":68,"repositoryHomepageUrl":69,"ownerHtmlUrl":70,"ownerBlogUrl":69,"ownerDescription":71,"updatedAt":72,"install":73,"resources":86},"01M6CZ592A5T2RP7CVHNFWK6HH","article",null,"Write a publication-ready article in one of three angles - a trending long-form piece, a watched-repo thesis, or a project-through-a-lens essay. Optional Replicate hero image with --visual.","aeonfun\u002Faeon","aeonfun","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F287715280?v=4","01M6CZ23C6MXGMF4N0V1DZ1247","aeon","01M6CZ23GMZP1BR424J8Y8K3DG",793,"MIT","https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon\u002Fblob\u002Fmain\u002Fskills\u002Farticle\u002FSKILL.md",[18,23,26],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000016","图像与多媒体","SCENARIO","使用场景",{"categoryNo":24,"categoryName":25,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000018","写作与文案",{"categoryNo":27,"categoryName":28,"taxonomyType":29,"taxonomyTypeDesc":30},"01JBSEEDCAT000000000000036","内容创作者","PROFESSION","适用职业",[5,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51],"basics","Publication","long","form","long-form","essay","hero","trending","lens","thesis","dev","repo","three","token","write","replicate","api","image","project","ready","2026-11-10 10:12:30",false,"---\nname: article\ndescription: Write a publication-ready article in one of three angles - a trending long-form piece, a watched-repo thesis, or a project-through-a-lens essay. Optional Replicate hero image with --visual.\nmetadata:\n  title: Article\n  category: basics\n  var: \"\"\n  tags:\n    - content\n    - dev\n  requires:\n    - REPLICATE_API_TOKEN?\n---\n> **${var}** — Selector: `[angle:arg] [--visual]`. The **angle** prefix picks the article type; append **`--visual`** (or `visual`) anywhere to also generate a Replicate hero image.\n>\n> - **empty** → `standard` general long-form article on an auto-selected trending topic. If the resolved topic is a single explainable mechanism, it becomes a technical explainer instead.\n> - **`\u003Ctopic>`** (no recognized prefix) → `standard` article on that topic.\n> - **`repo:\u003Cowner\u002Frepo>`** → `repo` thesis-driven article about that repo. `repo:\u003Cangle>` (e.g. `repo:architecture`) or bare `repo:` uses the repo from `memory\u002Fwatched-repos.md` with that angle \u002F an auto-selected angle — this preserves repo-article's original input.\n> - **`lens:\u003Ctopic>`** → `lens` project-through-a-lens essay framed by that lens (e.g. `lens:unix philosophy`). Bare `lens:` auto-selects the lens.\n> - **`--visual`** appended to any of the above → after the body is written, generate a Replicate hero image (optional `REPLICATE_API_TOKEN`; ships text-only if absent).\n>\n> Examples: `\"\"`, `\"entropy trajectory reasoning --visual\"`, `\"repo:aeonfun\u002Faeon\"`, `\"repo:roadmap\"`, `\"lens:regulation wave --visual\"`.\n\nToday is ${today}. Write a high-quality, publication-ready article. No placeholders.\n\n## Shared preamble (every run)\n\n1. Read `memory\u002FMEMORY.md` for context on what topics\u002Farticles have been covered recently.\n2. Read the last 3–7 days of `memory\u002Flogs\u002F` for recent activity — and **don't re-report** something already covered.\n3. **Parse `${var}` into `angle` + `visual`:**\n   - Detect a standalone `--visual` or `visual` token anywhere in `${var}`; if present set **`visual = true`** and strip that token. Otherwise `visual = false`.\n   - From what remains: if it starts with `repo:` → `angle = repo`, `arg =` the rest. If it starts with `lens:` → `angle = lens`, `arg =` the rest. Otherwise → `angle = standard`, `arg =` the whole remaining string (empty ⇒ auto-select).\n4. Dispatch to the matching angle section below. If `visual = true`, run the **Visual add-on** after the article body is written, regardless of angle.\n\n---\n\n## Angle: standard — long-form article \u002F technical explainer\n\nA single long-form article. It takes one of two structures depending on the topic:\n\n- **General article** — a broad trend, development, or event. 600–800 words.\n- **Technical explainer** — a single explainable mechanism, technique, algorithm, or system. 600–1000 words, using the explainer structure below.\n\n### Topic selection (standard)\n\n- If `arg` (the topic) is set, use it verbatim. If it clearly names a single mechanism\u002Ftechnique\u002Fsystem → **technical explainer** structure; otherwise → **general article** structure.\n- If `arg` is empty, pick deterministically — first hit wins:\n  1. **Explainer candidate:** a single most non-obvious mechanism inside the newest file in `output\u002Farticles\u002F` from the last 3 days; else the newest \"Paper Pick\" in `memory\u002Flogs\u002F` from the last 7 days (its headline mechanism); else a specific technique\u002Falgorithm\u002Fsystem surfaced in the last 7 days of logs. If a strong single-mechanism candidate exists → **technical explainer** on it. Reject any candidate broader than a single mechanism (e.g. \"AI agents\" — too vague; \"MCP tool-routing via vector search\" — usable).\n  2. **General candidate:** otherwise search the web for the most interesting recent development in AI, crypto\u002FDeFi, or consciousness research — pick whichever has the most compelling story today (WebSearch) → **general article**.\n\n### Voice (technical explainer)\n\nIf a `soul\u002F` directory exists, read the soul files for voice calibration: `soul\u002FSOUL.md` (identity, worldview, opinions), then `soul\u002FSTYLE.md` (writing style, sentence structure, anti-patterns). This is *you explaining a mechanism to a smart friend* — more precision than a general article, same voice. No textbook tone, no \"let's explore.\" If `soul\u002F` is empty, default to clear, direct, neutral.\n\n### Research\n\n**General article:** read 2–3 source articles with WebFetch to gather facts and quotes.\n\n**Technical explainer:** run **three distinct WebSearch queries** so you triangulate rather than echo one source:\n1. `\"\u003Ctopic>\" how it works` — mechanism explanations\n2. `\"\u003Ctopic>\" benchmark OR results OR latency OR cost` — concrete numbers\n3. `\"\u003Ctopic>\" limits OR criticism OR fails OR doesn't work` — failure modes and pushback\n\nIf the topic is from a paper, also fetch the paper metadata and abstract (one call, TOPIC written in literally):\n```bash\ncurl -s \"https:\u002F\u002Fapi.semanticscholar.org\u002Fgraph\u002Fv1\u002Fpaper\u002Fsearch?query=TOPIC&limit=5&fields=title,authors,abstract,url,publicationDate,openAccessPdf\"\n```\nIf the curl fails, use WebFetch on `https:\u002F\u002Fwww.semanticscholar.org\u002Fsearch?q=TOPIC` instead (see **Shell shape** in the Network note).\nUse **WebFetch** to read the 2–3 best sources in depth. **At least one source must be primary**: a paper (arXiv \u002F OpenReview \u002F Semantic Scholar), official documentation, the project's own README, or a code repo. Blog summaries alone are not enough — they often mangle the mechanism.\n\nExtract:\n- The **single core mechanism** — the one move that, once you grok it, makes the rest fall into place.\n- A **vivid analogy** for the mechanism, and the precise place where the analogy breaks down (the breakage is the interesting part).\n- **3–5 specific numbers** — benchmarks, latencies, costs, error rates, training compute, parameter counts. Each number gets a source URL.\n- **What would falsify this** — what result, if observed, would mean the mechanism doesn't work as claimed. If you can't name one, the explanation isn't sharp enough — keep digging.\n\n### Write\n\n**General article** — 600–800 words in Markdown. Include:\n- A compelling title\n- A short intro hook\n- 3–4 substantive sections\n- Cited sources (with URLs) at the bottom\n\n**Technical explainer** — 600–1000 words. Structure (every section required):\n```\n# \u003CTitle>\n\n**Key idea in one sentence:** \u003Cone-sentence claim about the mechanism>\n\n## The Setup\n2-3 sentences. What problem does this solve? Why now?\n\n## The Intuition Pump\nA vivid analogy that builds the reader's mental model in 3-4 sentences. Then one sentence on **where the analogy breaks down** - that's where the real mechanism lives.\n\n## How It Actually Works\nA numbered walkthrough of the mechanism in **3-7 steps**. Each step is one or two sentences. Use concrete examples - name the specific function, layer, message, opcode, contract. No \"the system processes the input\" - say what the system actually does.\n\n## Numbers That Anchor It\n3-5 bullet points. Each bullet is a specific number with a source link, e.g.:\n- 8.4× faster end-to-end than baseline at 4K context ([source](url))\n\n## What Would Break This\n1-2 sentences naming a result that, if observed, would falsify the claim. This forces honesty.\n\n## Why It Matters\n2-3 sentences. What does this unlock? Who should care?\n\n## Sources\n- [Title 1](url) - primary\n- [Title 2](url)\n- [Title 3](url)\n```\n\n**Voice rules (technical explainer):** First person where it fits. Explanatory > opinionated, but not bloodless. Technical precision > hedging — if you don't know, say so, don't fudge. Short paragraphs. Em dashes. Concrete > abstract. Reference specific systems, papers, people — no \"researchers have shown,\" name them. Cite inline: every number, every claim that could be wrong, gets a link.\n\n### Save & notify (standard)\n\n- **General article:** save to `output\u002Farticles\u002F${today}.md`.\n- **Technical explainer:** save to `output\u002Farticles\u002Fexplainer-${today}.md`. If a hero image was generated (see Visual add-on), put it at the very top: `![hero](..\u002Fimages\u002Fexplainer-${today}.\u003Cext>)` — relative path, skip the line if no image — and add an HTML comment with the image prompt used (for future audits).\n\nUpdate `memory\u002FMEMORY.md` to record the article and its topic (add to the `Recent Articles` list\u002Ftable). Append the consolidated log entry (see **Log**), then notify via `.\u002Fnotify`:\n\n- **General article:**\n  ```\n  New article written: [title]\n\n  https:\u002F\u002Fgithub.com\u002F\u003CGITHUB_REPOSITORY>\u002Fblob\u002Fmain\u002Foutput\u002Farticles\u002F${today}.md\n  ```\n  Use the `GITHUB_REPOSITORY` env var (GitHub Actions sets it to `owner\u002Frepo` of the running instance): print it with `gh repo view --json nameWithOwner -q .nameWithOwner` as its own call and write it in literally.\n\n- **Technical explainer:**\n  ```\n  technical explainer: [title]\n\n  [the one-sentence \"key idea\" line, verbatim]\n\n  [hero image URL if generated - original Replicate URL still works for ~24h]\n\n  read it: output\u002Farticles\u002Fexplainer-${today}.md\n  ```\n\n---\n\n## Angle: repo — thesis-driven article about a watched repo\n\n\u003C!-- autoresearch: variation B — editorial discipline: research → thesis → draft → self-edit, with a falsifiable claim and a quality gate -->\n\n### Config\n\nReads repos from `memory\u002Fwatched-repos.md`. Resolve the target repo:\n- If `arg` looks like `owner\u002Frepo` (contains a `\u002F`) → that's the repo to cover; the angle is auto-selected in Phase 2.\n- Else if `arg` is a non-empty keyword (e.g. `architecture`, `recent progress`, `roadmap`) → it's the **angle**; pick the repo from `memory\u002Fwatched-repos.md` (if multiple are listed, the one with the most activity in the last 7 days).\n- Else (`arg` empty) → repo from `memory\u002Fwatched-repos.md` (most active of the last 7 days), angle auto-selected.\n\nAn article without a thesis is filler. This angle runs five phases and only advances when the current phase's gate passes.\n\n### Phase 1 — Research (gather, don't write yet)\n\nRun these in parallel where possible (substitute the resolved `owner\u002Frepo`), each its own Bash call. First the 7-day cutoff:\n```bash\ndate -u -d '7 days ago' +%Y-%m-%dT%H:%M:%SZ\n```\nIt prints SINCE; write it in literally where `\u003CSINCE>` appears below.\n\nRepo metadata:\n```bash\ngh api repos\u002Fowner\u002Frepo --jq '{name, description, language, stargazers_count, forks_count, open_issues_count, topics, created_at, updated_at, pushed_at, default_branch}'\n```\n\nCommits in last 7 days (paginated):\n```bash\ngh api repos\u002Fowner\u002Frepo\u002Fcommits -X GET \\\n  -f since=\"\u003CSINCE>\" \\\n  --jq '.[] | {sha: .sha[0:7], msg: .commit.message | split(\"\\n\")[0], author: .commit.author.name, date: .commit.author.date, url: .html_url}' --paginate\n```\n\nMerged PRs in last 7 days:\n```bash\ngh api 'repos\u002Fowner\u002Frepo\u002Fpulls?state=closed&sort=updated&direction=desc&per_page=50' \\\n  --jq '[.[] | select(.merged_at and (.merged_at > (now - 86400*7 | todate))) | {number, title, user: .user.login, merged_at, additions, deletions, url: .html_url}]'\n```\n\nOpen PRs:\n```bash\ngh api repos\u002Fowner\u002Frepo\u002Fpulls --jq '[.[] | {number, title, user: .user.login, created_at, draft, labels: [.labels[].name], url: .html_url}]'\n```\n\nIssues opened\u002Fclosed in last 7 days (exclude PRs):\n```bash\ngh api 'repos\u002Fowner\u002Frepo\u002Fissues?state=all&since=\u003CSINCE>&per_page=100' --paginate \\\n  --jq '[.[] | select(.pull_request | not) | {number, title, state, created_at, closed_at, labels: [.labels[].name]}]'\n```\n\nLast 3 releases:\n```bash\ngh api repos\u002Fowner\u002Frepo\u002Freleases --jq '.[0:3] | .[] | {tag_name, name, published_at, body}'\n```\n\nREADME (fallback: WebFetch raw URL if base64 decode fails):\n```bash\ngh api repos\u002Fowner\u002Frepo\u002Freadme --jq '.content' | base64 -d\n```\n\nFrom the commit list, find the most-frequently-touched files. Read the top 2–3 of those with `gh api repos\u002Fowner\u002Frepo\u002Fcontents\u002F\u003Cpath>` plus any `CHANGELOG.md`, `ROADMAP.md`, or architecture docs.\n\n**External context** — three distinct WebSearch queries:\n1. `\"owner\u002Frepo\" site:news.ycombinator.com OR site:lobste.rs OR site:reddit.com`\n2. `\"owner\u002Frepo\" twitter OR x.com` (or the project name if distinctive)\n3. One query to anchor positioning against a comparable\u002Fcompeting project.\n\n**Gate 1 — enough story?** If **all** of the following hold, abort and notify `REPO_ARTICLE_SKIPPED: insufficient activity` (log reason, write no article):\n- \u003C3 commits in the last 7 days, AND\n- 0 merged PRs in the last 7 days, AND\n- no release in the last 30 days, AND\n- no external mentions surfaced in step 3.\n\n**Quiet-repo exception**: if the repo has historical importance but is currently slow (e.g. only 1–2 commits this week, no release), do **not** skip — instead narrow the article's focus to the single most substantive recent change (a specific commit, a contested issue thread, a roadmap update) and write a shorter piece around *that*. Prefer publishing a tight 600-word piece on one real change over skipping.\n\n### Phase 2 — Thesis\n\nWrite one **falsifiable claim** in ≤25 words. The claim must be disprovable by specific evidence — not a vibe.\n- Good: \"aeonfun\u002Faeon is pivoting from scheduled digests to reactive skill chains — 4 of 7 merged PRs this week added or consumed `output\u002F.chains\u002F*.md` contracts.\"\n- Bad: \"Aeon is an interesting agent framework.\" (not falsifiable)\n\nIf an angle is forced (from `arg`), the thesis must relate to it (e.g. angle `architecture` → an architectural claim). If no angle is forced, pick the one with the strongest evidence from: shipping velocity shift, architectural pivot, community growth inflection, roadmap commitment, deprecation\u002Fscope cut, performance or scale milestone.\n\n**Gate 2 — falsifiability.** Finish the sentence: \"This claim would be wrong if ____.\" If you can't complete it with something concrete and checkable, rewrite the thesis.\n\n### Phase 3 — Draft (600–900 words, Markdown)\n\n```markdown\n# [Title that asserts the thesis or a consequence of it - not \"A look at X\"]\n\n[1-paragraph hook, ≤80 words: lead with the thesis or a surprising number that sets it up.]\n\n## The claim\n> [The falsifiable thesis, verbatim, as a blockquote.]\n\n## Evidence\n[Two to four sub-paragraphs. Each MUST cite at least one specific commit SHA, PR#, file path, release tag, or external mention. Link the source inline.]\n\n## Counter-evidence \u002F what would change my mind\n[One paragraph. What recent signals argue against the thesis? Be honest. If genuinely nothing does, say so - but only after looking.]\n\n## Why it matters\n[One paragraph. Who benefits or loses if the thesis is true? Connect to an ecosystem trend, user need, or competing project.]\n\n---\n*Sources*\n- [Label](url)\n- [Label](url)\n[≥4 total, ≥1 in-repo (commit\u002FPR link) and ≥1 external (news\u002Fsocial\u002Fdoc).]\n```\n\n### Phase 4 — Self-edit (required)\n\nRun this checklist. Rewrite any line that fails. Target: 8\u002F8 passing.\n1. **Thesis visible in first 100 words?** If not, rewrite the hook.\n2. **Every section has ≥1 specific number, SHA, PR#, filename, or date?** (generic adjectives don't count)\n3. **Zero banned phrases** (see *Banned phrase lexicon* section below — check against that explicit list).\n4. **Counter-evidence is real** — not a strawman like \"some might say it's complex\".\n5. **Sources ≥4 links, ≥1 in-repo, ≥1 external.**\n6. **Title asserts something** (not \"A look at X\" \u002F \"Exploring Y\").\n7. **Word count in 600–900** (hard bounds — trim or expand).\n8. **No placeholder phrases** like \"[TBD]\", \"[link]\", \"[title]\".\n\nIf any item still fails after one rewrite pass, publish with status `REPO_ARTICLE_DEGRADED` and note which items failed in the log — don't hide it.\n\n### Phase 5 — Save, log, notify (repo)\n\n1. Save the article to `output\u002Farticles\u002Frepo-article-${today}.md`. (If a hero image was generated via the Visual add-on, put `![hero](..\u002Fimages\u002Frepo-article-${today}.\u003Cext>)` at the top.)\n2. Append the consolidated log entry (see **Log**) **before** notifying.\n3. Update the `Recent Articles` table in `memory\u002FMEMORY.md` (Date | Title | Topic).\n4. Notify via `.\u002Fnotify`:\n   ```\n   *[Article title]*\n\n   Thesis: [one sentence]\n\n   Read: [link to output\u002Farticles\u002Frepo-article-${today}.md in THIS repo - get the repo name from `git remote get-url origin`, not the watched repo]\n   ```\n\n### Banned phrase lexicon (repo angle)\n\nReject a draft that contains any of these. Match case-insensitively, whole phrase or obvious variant:\n- \"in today's fast-paced world\"\n- \"leveraging\" \u002F \"leverage\" (as a verb meaning \"use\")\n- \"robust\"\n- \"game-changer\" \u002F \"game-changing\"\n- \"under the hood\" (unless the section actually walks through internals)\n- \"taking X to the next level\"\n- \"at the end of the day\"\n- \"diving into\" \u002F \"deep dive\"\n- \"delving into\" \u002F \"delve\"\n- \"comprehensive suite\"\n- \"cutting-edge\"\n- \"seamlessly\" \u002F \"seamless\"\n- \"empowers\" \u002F \"empowering\"\n- \"revolutionize\" \u002F \"revolutionary\"\n- \"unlock\" (metaphorical, e.g. \"unlocks new possibilities\")\n- \"streamline\" (as filler)\n- \"best-in-class\"\n- \"paradigm shift\"\n\nIf a banned phrase is the *most accurate* word in a technical context (e.g. actually describing leverage in a derivatives article), keep it and note the exemption in the log.\n\n### Constraints (repo angle)\n\n- Never publish without a thesis.\n- Never pad to hit word count — 600 honest words beat 900 padded.\n- Never fabricate a SHA, PR number, or quote. If real evidence isn't available, weaken the thesis or skip.\n\n---\n\n## Angle: lens — the project through a surprising lens\n\n\u003C!-- autoresearch: variation B — editorial discipline (research → falsifiable thesis → draft → self-edit with hard gates) -->\n\nWrites articles that explain the project through a **different lens each time** — framed so a reader who's never heard of the project understands why it matters, via something they already care about. NOT a repo progress update (that's the **repo** angle above). `arg` is the lens (e.g. \"unix philosophy\", \"regulation wave\", \"open source funding\"); if empty, auto-select from trending topics + angle rotation.\n\nRead before deciding anything: `memory\u002FMEMORY.md`, the last 7 days of `memory\u002Flogs\u002F`, `memory\u002Fwatched-repos.md`, and `memory\u002Fproject-lens-angles.md` (may not exist on first run — treat absence as empty history).\n\n**Why models fail at this by default:** they slide into feature-listing wrapped in philosophical language, forced parallels with no mechanism, and marketing tone. This angle prevents that with a research → thesis → draft → self-edit pipeline where each phase has hard gates. If the gates can't pass, abort — don't publish a weak article.\n\n### Phase 1 — Context\n\nRead before deciding anything:\n- Last 14 days of `output\u002Farticles\u002Fproject-lens-*.md` and `memory\u002Fproject-lens-angles.md` — know which angle categories and theses are exhausted.\n- 2–3 most recent `output\u002Farticles\u002Frepo-article-*.md` and `output\u002Farticles\u002Fpush-recap-*.md` — know what shipped lately.\n- Repo state: `gh api repos\u002F{owner}\u002F{repo} --jq '{name, description, stargazers_count, forks_count, open_issues_count, updated_at}'`. If unreachable, continue with memory only and log the gap.\n\nIf `memory\u002Fwatched-repos.md` is empty or missing, abort and notify: \"project-lens: no watched repo configured.\"\n\n### Phase 2 — Pick the lens\n\n**If `arg` is set**, use it verbatim. Classify into one of the 8 categories below for logging.\n\n**If `arg` is empty**:\n1. Run 2–3 WebSearch queries on what's being debated right now in tech, crypto, AI, regulation, open source, or philosophy (e.g., `\"AI agents\" autonomy debate last 7 days`, `crypto regulation April 2026`, `open source funding model 2026`).\n2. From results, identify 3 candidate angles with non-obvious connections to the project.\n3. Pick the one that (a) hasn't appeared in the last 14 days **and** (b) has the strongest concrete connection. Record the choice and the rejected candidates with one-line reasons.\n\n**Angle categories (no repeat within 14 days):**\n1. **Current events** — Something happening this week\u002Fmonth.\n2. **Philosophy \u002F big ideas** — Unix philosophy, cathedral vs bazaar, composability, anti-fragility, skin in the game, swarm intelligence, etc.\n3. **Industry comparison** — How a well-known company\u002Fproject solved a similar problem differently.\n4. **User story** — POV of a specific persona (solo dev, DAO, research lab, crypto community) with and without this tool.\n5. **Contrarian take** — Challenge a common assumption; use project as evidence.\n6. **Technical deep-dive for non-technical readers** — One architectural decision, plain language, bigger implications.\n7. **Historical parallel** — Computing \u002F internet \u002F non-tech history with a concrete mechanism (not surface resemblance).\n8. **Ecosystem map** — Where the project sits: adjacent, complementary, competing.\n\n### Phase 3 — Research (gate: collect evidence before drafting)\n\n**External side — required minimums:**\n- 3+ WebSearch queries on the lens topic (different framings, not rewordings)\n- WebFetch on the 2+ most relevant sources\n- ≥3 **distinct domains** across cited sources\n- ≥3 concrete facts extracted: names, numbers, dollar amounts, dated quotes, specific events\n- Recency: ≤30 days old for \"current events\"; ≤180 days for industry comparison \u002F contrarian \u002F ecosystem; ≤5 years for philosophy \u002F historical\n- Log every URL consulted\n\n**Project side — required minimums:**\n- 2+ recent articles in `output\u002Farticles\u002F` read end-to-end\n- `gh api repos\u002F{owner}\u002F{repo}\u002Fcommits --jq '.[0:10] | .[] | {sha: .sha[0:7], msg: (.commit.message|split(\"\\n\")[0])}'` — last 10 commits\n- ≥3 specific project references you plan to use: named features, file paths, commit hashes, architectural choices. **Not** vague claims like \"the project uses AI\" or \"it has good UX.\"\n\n**If you cannot hit these minimums, abandon the angle and re-run Phase 2 with a different category.** Log the abandoned angle and why.\n\n### Phase 4 — Thesis lock (hard gate)\n\nBefore drafting, write **ONE falsifiable claim in ≤30 words** that links the lens to the project. Example:\n> \"Running agents as scheduled GitHub Actions — rather than as persistent servers — trades a few seconds of latency for a property the AI industry barely has: versioned, audit-trailed, publicly forkable autonomy.\"\n\nRules:\n- **Falsifiable**: a reasonable critic could argue the opposite.\n- **Specific**: names concrete things (cron jobs, not \"infrastructure\"; audit-trailed, not \"better\").\n- Not a tautology. Not marketing. Not \"this is cool because X.\"\n\n**If you can't state the thesis in one sentence, the angle isn't working — return to Phase 2.** Do not proceed with a fuzzy thesis.\n\n### Phase 5 — Draft (700–1000 words)\n\nSave to `output\u002Farticles\u002Fproject-lens-${today}.md` with this structure:\n```markdown\n# [Title: leads with the lens, works for a reader who doesn't know the project]\n\n[¶1-2: external hook. Start with the trend\u002Fidea\u002Fevent\u002Fquestion the reader already cares about. Do NOT name the project yet.]\n\n## [Section: establishes the external frame]\n[Build the lens with one or more of your concrete facts - a quote, a number, a specific event.]\n\n## [Section: introduces the project through the frame]\n[Project enters here - but through the lens, not as a feature list. Describe how it embodies, challenges, or extends the idea with specific code\u002Fdesign references.]\n\n## [Section: one non-obvious technical or strategic detail]\n[Where the article earns its existence. Point to something in the code, architecture, or approach a reader wouldn't get from the README.]\n\n## [Section: zoom back out]\n[A concrete forward claim - specific enough to be wrong. Not \"this is exciting.\" Something like \"this suggests X won't happen for 2-3 years because Y\" or \"this is the same mistake [named case] made, and it took [duration] to recover.\"]\n\n---\n*Sources:*\n- [source title](url) - what it was used for\n- ...\n```\n\n**Draft requirements:**\n- Title must work for a reader who doesn't know the project name.\n- ≥3 external citations rendered as inline links.\n- ≥3 specific project references (named features, file paths, commit hashes, named decisions).\n- 700–1000 words (count before submitting).\n\n### Phase 6 — Self-edit (hard gates — all must pass)\n\nGo through this checklist after the first draft. If any gate fails, rewrite the affected section **once**. If the second pass still fails, **abort and log** — do not publish a weak article.\n- [ ] Title does NOT name the project\n- [ ] First 2 paragraphs do NOT name the project\n- [ ] ≥3 external citations with URLs (inline)\n- [ ] ≥3 specific project references (named, not vague)\n- [ ] Falsifiable thesis visible in the article text\n- [ ] 700–1000 words\n- [ ] No banned phrases: *revolutionary*, *groundbreaking*, *game-changing*, *paradigm shift*, *disrupting*, *unlocks*, *empowers*, *the future of X*, *leverage \u002F leveraging*, *at scale*, *democratize* (unless quoting a source that used the word)\n- [ ] Every parallel\u002Fcomparison states a concrete mechanism — not surface resemblance\n- [ ] Closing section makes a specific forward claim (not generic optimism or \"time will tell\")\n- [ ] No feature-list paragraphs (\"the project does X, Y, Z\") — if found, cut and keep ONE element\n\n### Phase 7 — Output (lens)\n\n1. **Save** `output\u002Farticles\u002Fproject-lens-${today}.md`. (If a hero image was generated via the Visual add-on, put `![hero](..\u002Fimages\u002Fproject-lens-${today}.\u003Cext>)` at the top.)\n2. **Append** to `memory\u002Fproject-lens-angles.md` (create if missing):\n   ```markdown\n   ## ${today}\n   - Angle: [category]\n   - Thesis: [one-line falsifiable claim]\n   - Title: [article title]\n   - Sources: [3-5 URLs]\n   ```\n3. **Notify** via `.\u002Fnotify`:\n   ```\n   *New Article: [title]*\n\n   [3-4 sentence summary: the external thing the article connects to, the thesis claim, one specific project detail.]\n\n   Read: [URL to output\u002Farticles\u002Fproject-lens-${today}.md - use `git remote get-url origin` for this repo]\n   ```\n4. **Log** the consolidated entry (see **Log**).\n\n### Anti-patterns (prevention beats self-edit catch)\n\n- **Forced parallels** — every comparison needs a concrete mechanism, not surface resemblance. \"X is like Y because both are new\" fails; \"X is like Y because both decoupled [specific function] from [specific bottleneck]\" passes.\n- **Feature-dump via the lens** — pick ONE architectural decision and interrogate it, don't list what the project does.\n- **Marketing tone** — aim for trade-publication prose, not a company blog.\n- **False novelty** — if you can't point to what's actually new, name what's old that still works and why.\n- **Vague closings** — the final section must make a claim specific enough that a reader could come back in six months and say \"you were wrong\" or \"you were right.\"\n\n### Constraints (lens angle)\n\n- Never publish if Phase 6 self-edit fails twice — abort cleanly.\n- Never reuse an angle category within 14 days (check `memory\u002Fproject-lens-angles.md`).\n- Never invent facts to fill the citation minimum — if research is thin, abandon the angle.\n\n---\n\n## Visual add-on (`--visual`) — Replicate hero image\n\nRuns **only when `visual = true`**, after the article body is written and saved, for **any** angle. Use Replicate's Nano Banana Pro (Gemini 3 Pro Image). It renders **text labels well** — exploit that by writing prompts that ask for labeled diagrams or schematics, not stock-photo metaphors. Set `IMG_BASENAME` to match the article file for the angle: `explainer-${today}` (standard\u002Fexplainer), `article-${today}` (standard\u002Fgeneral), `repo-article-${today}` (repo), or `project-lens-${today}` (lens).\n\n1. **Preflight**: read the \"Keys set for this run\" line in your run notes, never a shell test (those are refused and cost a turn). If `REPLICATE_API_TOKEN` is not on that line, log `IMAGE_SKIPPED reason=no-token` and skip to step 5 (no-image path). Do not attempt any Replicate call. The article must ship without an image in this case.\n\n2. **Craft the prompt**. Aim for technical illustration energy, not marketing. Strong prompt templates:\n   - *Schematic*: \"Technical schematic illustration of \u003Cmechanism>, dark navy background, thin cyan and amber lines, labeled boxes reading '\u003Clabel1>', '\u003Clabel2>', '\u003Clabel3>', arrows showing data flow from \u003CA> to \u003CB> to \u003CC>, blueprint aesthetic, 16:9\"\n   - *Conceptual*: \"Editorial illustration capturing \u003Ccore concept>: \u003Cvisual metaphor with concrete objects>, flat geometric style, restrained palette of two accent colors on near-black background, no human figures, 16:9\"\n   - *Data-flow*: \"Network diagram of \u003Cmechanism>: nodes labeled '\u003CA>', '\u003CB>', '\u003CC>' connected by directional arrows, weights shown as line thickness, monospace labels, technical-paper figure style, 16:9\"\n   Avoid: photorealistic faces, stock-business imagery, \"AI brain\" tropes, gradient slop.\n\n3. **Generate** with fallback enabled from the start (Nano Banana Pro can rate-limit; Seedream 5.0 lite is the fallback). The Replicate call is auth'd, so route it through `.\u002Fsecretcurl` with the `{REPLICATE_API_TOKEN}` placeholder — never a bare `$REPLICATE_API_TOKEN` on the line (the Bash permission layer refuses it):\n   ```bash\n   .\u002Fsecretcurl -s -X POST \\\n     -H \"Authorization: Bearer {REPLICATE_API_TOKEN}\" \\\n     -H \"Content-Type: application\u002Fjson\" \\\n     -H \"Prefer: wait\" \\\n     -d '{\n       \"input\": {\n         \"prompt\": \"YOUR_DETAILED_PROMPT_HERE\",\n         \"aspect_ratio\": \"16:9\",\n         \"number_of_images\": 1,\n         \"safety_tolerance\": 5,\n         \"allow_fallback_model\": true\n       }\n     }' \\\n     \"https:\u002F\u002Fapi.replicate.com\u002Fv1\u002Fmodels\u002Fgoogle\u002Fnano-banana-pro\u002Fpredictions\"\n   ```\n   `Prefer: wait` usually returns the image inline in `.output`. If `.output` is empty, the prediction is still running - poll `.urls.get` for up to ~60s (`.\u002Fsecretcurl -s -H \"Authorization: Bearer {REPLICATE_API_TOKEN}\" \"\u003CPRED_URL>\"`, the `.urls.get` value written in literally), stopping when `.status` is `succeeded` (read `.output`) or `failed`\u002F`canceled` (no-image path, step 5).\n\n4. **Persist locally** — Replicate CDN URLs expire. Download and commit (the CDN URL carries no secret, so plain `curl` is fine):\n   ```bash\n   mkdir -p output\u002Fimages\n   ```\n   `IMAGE_URL` = the URL extracted from `response.output`. `EXT` = the last `.jpg`, `.jpeg`, `.png` or `.webp` in IMAGE_URL, or `.jpg` if it has none (work it out yourself). `LOCAL_PATH` = `output\u002Fimages\u002F\u003CIMG_BASENAME>\u003CEXT>`. Then download with every value written in literally:\n   ```bash\n   curl -sL \"\u003CIMAGE_URL>\" -o \"\u003CLOCAL_PATH>\"\n   ```\n   If the curl fails, retry via WebFetch or skip (no-image path, step 5).\n\n5. **No-image path** (token missing, API down, rate-limited, or download failed): log `IMAGE_SKIPPED reason=\u003Cconcrete reason>` and proceed with the article. Add a one-line note at the top of the article: `\u003C!-- hero image skipped: \u003Creason> -->`. The text must stand on its own. Never fail the whole skill because of an image problem.\n\nOnce the image is saved, add the hero-image line to the top of the article file (`![hero](..\u002Fimages\u002F${IMG_BASENAME}.\u003Cext>)`) and include the original Replicate URL in that angle's notification if the angle's notify format has an image slot.\n\n---\n\n## Log\n\nAppend **one** entry under a single `### article` heading in `memory\u002Flogs\u002F${today}.md`, as bullet points. Start with a discriminator line naming the branch\u002Fmode that ran, then the branch-specific fields:\n\n```\n### article\n- Branch: standard | repo | lens   (+visual if --visual ran)\n```\n\n**Standard branch fields:**\n```\n- Mode: general-article | technical-explainer\n- Topic: [topic]\n- Title: [title]\n- Key idea: [one-sentence claim]   (technical-explainer only)\n- Image: generated | fallback-model | skipped (\u003Creason>) | n\u002Fa\n- Image prompt: [prompt used, or \"n\u002Fa\"]\n- Primary source: [URL]            (technical-explainer only)\n- File: output\u002Farticles\u002F${today}.md | output\u002Farticles\u002Fexplainer-${today}.md\n- Notification sent: yes | no\n```\n\n**Repo branch fields:**\n```\n- Repo: owner\u002Frepo\n- Thesis: [verbatim]\n- Angle: [arg or auto-selected]\n- Word count: N\n- Self-edit checklist: X\u002F8 passing\n- Image: generated | fallback-model | skipped (\u003Creason>) | n\u002Fa\n- Status: REPO_ARTICLE_OK | REPO_ARTICLE_DEGRADED | REPO_ARTICLE_SKIPPED\n```\n\n**Lens branch fields:**\n```\n- Angle: [category]\n- Thesis: [one-line]\n- External sources: [count] across [N] distinct domains\n- Project references: [count]\n- Self-edit gates: all passed | failed at [gate name] → rewrite → [passed | aborted]\n- Image: generated | fallback-model | skipped (\u003Creason>) | n\u002Fa\n- Status: published | aborted\n- Notification: sent | skipped\n```\n\nLog **always — even on partial failure** (e.g. IMAGE_SKIPPED, REPO_ARTICLE_SKIPPED, aborted lens).\n\n## Network note\n\nThere is no network sandbox — `curl` works. For a flaky public GET, fall back to **WebFetch** on the same URL. For an auth'd API, call `.\u002Fsecretcurl` with a `{ENV_NAME}` placeholder (the key is injected via `requires:`) — never a bare `$SECRET` on the line. `gh api` handles GitHub auth internally — prefer it over raw curl for repo metadata.\n\n**Shell shape.** Each Bash call is one allowed command (pipes are fine). The run refuses a call that sets a variable (`EXT=$(...)`, `IMAGE_URL=...`), contains a `$(...)` substitution, uses a `[ ... ]` test, or joins commands with `;`, `&&` or `||` (including an `|| echo` fallback), and each refusal costs a model turn. So every command in this skill is its own call: a `date` lookup prints a value you write into later commands literally (`\u003CSINCE>`, `\u003CIMAGE_URL>`, `\u003CEXT>`), the token preflight reads the run notes' \"Keys set for this run\" line instead of a shell test, and a fallback is a prose step you take when the call fails.\n\nThe Replicate call runs **in-run** via `.\u002Fsecretcurl` (see the Visual add-on). If it fails, times out, or the download fails, go straight to the no-image path (step 5) — the article ships text-only. There is no deferred fallback; the image is best-effort and never blocks the article.\n\n## Environment Variables\n- `REPLICATE_API_TOKEN` — Replicate API key, used only by the `--visual` add-on. Optional: article text ships without it via the no-image path.\n\nWrite complete, publication-ready content. No placeholders.\n","> **${var}** — Selector: `[angle:arg] [--visual]`. The **angle** prefix picks the article type; append **`--visual`** (or `visual`) anywhere to also generate a Replicate hero image.\n>\n> - **empty** → `standard` general long-form article on an auto-selected trending topic. If the resolved topic is a single explainable mechanism, it becomes a technical explainer instead.\n> - **`\u003Ctopic>`** (no recognized prefix) → `standard` article on that topic.\n> - **`repo:\u003Cowner\u002Frepo>`** → `repo` thesis-driven article about that repo. `repo:\u003Cangle>` (e.g. `repo:architecture`) or bare `repo:` uses the repo from `memory\u002Fwatched-repos.md` with that angle \u002F an auto-selected angle — this preserves repo-article's original input.\n> - **`lens:\u003Ctopic>`** → `lens` project-through-a-lens essay framed by that lens (e.g. `lens:unix philosophy`). Bare `lens:` auto-selects the lens.\n> - **`--visual`** appended to any of the above → after the body is written, generate a Replicate hero image (optional `REPLICATE_API_TOKEN`; ships text-only if absent).\n>\n> Examples: `\"\"`, `\"entropy trajectory reasoning --visual\"`, `\"repo:aeonfun\u002Faeon\"`, `\"repo:roadmap\"`, `\"lens:regulation wave --visual\"`.\n\nToday is ${today}. Write a high-quality, publication-ready article. No placeholders.\n\n## Shared preamble (every run)\n\n1. Read `memory\u002FMEMORY.md` for context on what topics\u002Farticles have been covered recently.\n2. Read the last 3–7 days of `memory\u002Flogs\u002F` for recent activity — and **don't re-report** something already covered.\n3. **Parse `${var}` into `angle` + `visual`:**\n   - Detect a standalone `--visual` or `visual` token anywhere in `${var}`; if present set **`visual = true`** and strip that token. Otherwise `visual = false`.\n   - From what remains: if it starts with `repo:` → `angle = repo`, `arg =` the rest. If it starts with `lens:` → `angle = lens`, `arg =` the rest. Otherwise → `angle = standard`, `arg =` the whole remaining string (empty ⇒ auto-select).\n4. Dispatch to the matching angle section below. If `visual = true`, run the **Visual add-on** after the article body is written, regardless of angle.\n\n---\n\n## Angle: standard — long-form article \u002F technical explainer\n\nA single long-form article. It takes one of two structures depending on the topic:\n\n- **General article** — a broad trend, development, or event. 600–800 words.\n- **Technical explainer** — a single explainable mechanism, technique, algorithm, or system. 600–1000 words, using the explainer structure below.\n\n### Topic selection (standard)\n\n- If `arg` (the topic) is set, use it verbatim. If it clearly names a single mechanism\u002Ftechnique\u002Fsystem → **technical explainer** structure; otherwise → **general article** structure.\n- If `arg` is empty, pick deterministically — first hit wins:\n  1. **Explainer candidate:** a single most non-obvious mechanism inside the newest file in `output\u002Farticles\u002F` from the last 3 days; else the newest \"Paper Pick\" in `memory\u002Flogs\u002F` from the last 7 days (its headline mechanism); else a specific technique\u002Falgorithm\u002Fsystem surfaced in the last 7 days of logs. If a strong single-mechanism candidate exists → **technical explainer** on it. Reject any candidate broader than a single mechanism (e.g. \"AI agents\" — too vague; \"MCP tool-routing via vector search\" — usable).\n  2. **General candidate:** otherwise search the web for the most interesting recent development in AI, crypto\u002FDeFi, or consciousness research — pick whichever has the most compelling story today (WebSearch) → **general article**.\n\n### Voice (technical explainer)\n\nIf a `soul\u002F` directory exists, read the soul files for voice calibration: `soul\u002FSOUL.md` (identity, worldview, opinions), then `soul\u002FSTYLE.md` (writing style, sentence structure, anti-patterns). This is *you explaining a mechanism to a smart friend* — more precision than a general article, same voice. No textbook tone, no \"let's explore.\" If `soul\u002F` is empty, default to clear, direct, neutral.\n\n### Research\n\n**General article:** read 2–3 source articles with WebFetch to gather facts and quotes.\n\n**Technical explainer:** run **three distinct WebSearch queries** so you triangulate rather than echo one source:\n1. `\"\u003Ctopic>\" how it works` — mechanism explanations\n2. `\"\u003Ctopic>\" benchmark OR results OR latency OR cost` — concrete numbers\n3. `\"\u003Ctopic>\" limits OR criticism OR fails OR doesn't work` — failure modes and pushback\n\nIf the topic is from a paper, also fetch the paper metadata and abstract (one call, TOPIC written in literally):\n```bash\ncurl -s \"https:\u002F\u002Fapi.semanticscholar.org\u002Fgraph\u002Fv1\u002Fpaper\u002Fsearch?query=TOPIC&limit=5&fields=title,authors,abstract,url,publicationDate,openAccessPdf\"\n```\nIf the curl fails, use WebFetch on `https:\u002F\u002Fwww.semanticscholar.org\u002Fsearch?q=TOPIC` instead (see **Shell shape** in the Network note).\nUse **WebFetch** to read the 2–3 best sources in depth. **At least one source must be primary**: a paper (arXiv \u002F OpenReview \u002F Semantic Scholar), official documentation, the project's own README, or a code repo. Blog summaries alone are not enough — they often mangle the mechanism.\n\nExtract:\n- The **single core mechanism** — the one move that, once you grok it, makes the rest fall into place.\n- A **vivid analogy** for the mechanism, and the precise place where the analogy breaks down (the breakage is the interesting part).\n- **3–5 specific numbers** — benchmarks, latencies, costs, error rates, training compute, parameter counts. Each number gets a source URL.\n- **What would falsify this** — what result, if observed, would mean the mechanism doesn't work as claimed. If you can't name one, the explanation isn't sharp enough — keep digging.\n\n### Write\n\n**General article** — 600–800 words in Markdown. Include:\n- A compelling title\n- A short intro hook\n- 3–4 substantive sections\n- Cited sources (with URLs) at the bottom\n\n**Technical explainer** — 600–1000 words. Structure (every section required):\n```\n# \u003CTitle>\n\n**Key idea in one sentence:** \u003Cone-sentence claim about the mechanism>\n\n## The Setup\n2-3 sentences. What problem does this solve? Why now?\n\n## The Intuition Pump\nA vivid analogy that builds the reader's mental model in 3-4 sentences. Then one sentence on **where the analogy breaks down** - that's where the real mechanism lives.\n\n## How It Actually Works\nA numbered walkthrough of the mechanism in **3-7 steps**. Each step is one or two sentences. Use concrete examples - name the specific function, layer, message, opcode, contract. No \"the system processes the input\" - say what the system actually does.\n\n## Numbers That Anchor It\n3-5 bullet points. Each bullet is a specific number with a source link, e.g.:\n- 8.4× faster end-to-end than baseline at 4K context ([source](url))\n\n## What Would Break This\n1-2 sentences naming a result that, if observed, would falsify the claim. This forces honesty.\n\n## Why It Matters\n2-3 sentences. What does this unlock? Who should care?\n\n## Sources\n- [Title 1](url) - primary\n- [Title 2](url)\n- [Title 3](url)\n```\n\n**Voice rules (technical explainer):** First person where it fits. Explanatory > opinionated, but not bloodless. Technical precision > hedging — if you don't know, say so, don't fudge. Short paragraphs. Em dashes. Concrete > abstract. Reference specific systems, papers, people — no \"researchers have shown,\" name them. Cite inline: every number, every claim that could be wrong, gets a link.\n\n### Save & notify (standard)\n\n- **General article:** save to `output\u002Farticles\u002F${today}.md`.\n- **Technical explainer:** save to `output\u002Farticles\u002Fexplainer-${today}.md`. If a hero image was generated (see Visual add-on), put it at the very top: `![hero](..\u002Fimages\u002Fexplainer-${today}.\u003Cext>)` — relative path, skip the line if no image — and add an HTML comment with the image prompt used (for future audits).\n\nUpdate `memory\u002FMEMORY.md` to record the article and its topic (add to the `Recent Articles` list\u002Ftable). Append the consolidated log entry (see **Log**), then notify via `.\u002Fnotify`:\n\n- **General article:**\n  ```\n  New article written: [title]\n\n  https:\u002F\u002Fgithub.com\u002F\u003CGITHUB_REPOSITORY>\u002Fblob\u002Fmain\u002Foutput\u002Farticles\u002F${today}.md\n  ```\n  Use the `GITHUB_REPOSITORY` env var (GitHub Actions sets it to `owner\u002Frepo` of the running instance): print it with `gh repo view --json nameWithOwner -q .nameWithOwner` as its own call and write it in literally.\n\n- **Technical explainer:**\n  ```\n  technical explainer: [title]\n\n  [the one-sentence \"key idea\" line, verbatim]\n\n  [hero image URL if generated - original Replicate URL still works for ~24h]\n\n  read it: output\u002Farticles\u002Fexplainer-${today}.md\n  ```\n\n---\n\n## Angle: repo — thesis-driven article about a watched repo\n\n\u003C!-- autoresearch: variation B — editorial discipline: research → thesis → draft → self-edit, with a falsifiable claim and a quality gate -->\n\n### Config\n\nReads repos from `memory\u002Fwatched-repos.md`. Resolve the target repo:\n- If `arg` looks like `owner\u002Frepo` (contains a `\u002F`) → that's the repo to cover; the angle is auto-selected in Phase 2.\n- Else if `arg` is a non-empty keyword (e.g. `architecture`, `recent progress`, `roadmap`) → it's the **angle**; pick the repo from `memory\u002Fwatched-repos.md` (if multiple are listed, the one with the most activity in the last 7 days).\n- Else (`arg` empty) → repo from `memory\u002Fwatched-repos.md` (most active of the last 7 days), angle auto-selected.\n\nAn article without a thesis is filler. This angle runs five phases and only advances when the current phase's gate passes.\n\n### Phase 1 — Research (gather, don't write yet)\n\nRun these in parallel where possible (substitute the resolved `owner\u002Frepo`), each its own Bash call. First the 7-day cutoff:\n```bash\ndate -u -d '7 days ago' +%Y-%m-%dT%H:%M:%SZ\n```\nIt prints SINCE; write it in literally where `\u003CSINCE>` appears below.\n\nRepo metadata:\n```bash\ngh api repos\u002Fowner\u002Frepo --jq '{name, description, language, stargazers_count, forks_count, open_issues_count, topics, created_at, updated_at, pushed_at, default_branch}'\n```\n\nCommits in last 7 days (paginated):\n```bash\ngh api repos\u002Fowner\u002Frepo\u002Fcommits -X GET \\\n  -f since=\"\u003CSINCE>\" \\\n  --jq '.[] | {sha: .sha[0:7], msg: .commit.message | split(\"\\n\")[0], author: .commit.author.name, date: .commit.author.date, url: .html_url}' --paginate\n```\n\nMerged PRs in last 7 days:\n```bash\ngh api 'repos\u002Fowner\u002Frepo\u002Fpulls?state=closed&sort=updated&direction=desc&per_page=50' \\\n  --jq '[.[] | select(.merged_at and (.merged_at > (now - 86400*7 | todate))) | {number, title, user: .user.login, merged_at, additions, deletions, url: .html_url}]'\n```\n\nOpen PRs:\n```bash\ngh api repos\u002Fowner\u002Frepo\u002Fpulls --jq '[.[] | {number, title, user: .user.login, created_at, draft, labels: [.labels[].name], url: .html_url}]'\n```\n\nIssues opened\u002Fclosed in last 7 days (exclude PRs):\n```bash\ngh api 'repos\u002Fowner\u002Frepo\u002Fissues?state=all&since=\u003CSINCE>&per_page=100' --paginate \\\n  --jq '[.[] | select(.pull_request | not) | {number, title, state, created_at, closed_at, labels: [.labels[].name]}]'\n```\n\nLast 3 releases:\n```bash\ngh api repos\u002Fowner\u002Frepo\u002Freleases --jq '.[0:3] | .[] | {tag_name, name, published_at, body}'\n```\n\nREADME (fallback: WebFetch raw URL if base64 decode fails):\n```bash\ngh api repos\u002Fowner\u002Frepo\u002Freadme --jq '.content' | base64 -d\n```\n\nFrom the commit list, find the most-frequently-touched files. Read the top 2–3 of those with `gh api repos\u002Fowner\u002Frepo\u002Fcontents\u002F\u003Cpath>` plus any `CHANGELOG.md`, `ROADMAP.md`, or architecture docs.\n\n**External context** — three distinct WebSearch queries:\n1. `\"owner\u002Frepo\" site:news.ycombinator.com OR site:lobste.rs OR site:reddit.com`\n2. `\"owner\u002Frepo\" twitter OR x.com` (or the project name if distinctive)\n3. One query to anchor positioning against a comparable\u002Fcompeting project.\n\n**Gate 1 — enough story?** If **all** of the following hold, abort and notify `REPO_ARTICLE_SKIPPED: insufficient activity` (log reason, write no article):\n- \u003C3 commits in the last 7 days, AND\n- 0 merged PRs in the last 7 days, AND\n- no release in the last 30 days, AND\n- no external mentions surfaced in step 3.\n\n**Quiet-repo exception**: if the repo has historical importance but is currently slow (e.g. only 1–2 commits this week, no release), do **not** skip — instead narrow the article's focus to the single most substantive recent change (a specific commit, a contested issue thread, a roadmap update) and write a shorter piece around *that*. Prefer publishing a tight 600-word piece on one real change over skipping.\n\n### Phase 2 — Thesis\n\nWrite one **falsifiable claim** in ≤25 words. The claim must be disprovable by specific evidence — not a vibe.\n- Good: \"aeonfun\u002Faeon is pivoting from scheduled digests to reactive skill chains — 4 of 7 merged PRs this week added or consumed `output\u002F.chains\u002F*.md` contracts.\"\n- Bad: \"Aeon is an interesting agent framework.\" (not falsifiable)\n\nIf an angle is forced (from `arg`), the thesis must relate to it (e.g. angle `architecture` → an architectural claim). If no angle is forced, pick the one with the strongest evidence from: shipping velocity shift, architectural pivot, community growth inflection, roadmap commitment, deprecation\u002Fscope cut, performance or scale milestone.\n\n**Gate 2 — falsifiability.** Finish the sentence: \"This claim would be wrong if ____.\" If you can't complete it with something concrete and checkable, rewrite the thesis.\n\n### Phase 3 — Draft (600–900 words, Markdown)\n\n```markdown\n# [Title that asserts the thesis or a consequence of it - not \"A look at X\"]\n\n[1-paragraph hook, ≤80 words: lead with the thesis or a surprising number that sets it up.]\n\n## The claim\n> [The falsifiable thesis, verbatim, as a blockquote.]\n\n## Evidence\n[Two to four sub-paragraphs. Each MUST cite at least one specific commit SHA, PR#, file path, release tag, or external mention. Link the source inline.]\n\n## Counter-evidence \u002F what would change my mind\n[One paragraph. What recent signals argue against the thesis? Be honest. If genuinely nothing does, say so - but only after looking.]\n\n## Why it matters\n[One paragraph. Who benefits or loses if the thesis is true? Connect to an ecosystem trend, user need, or competing project.]\n\n---\n*Sources*\n- [Label](url)\n- [Label](url)\n[≥4 total, ≥1 in-repo (commit\u002FPR link) and ≥1 external (news\u002Fsocial\u002Fdoc).]\n```\n\n### Phase 4 — Self-edit (required)\n\nRun this checklist. Rewrite any line that fails. Target: 8\u002F8 passing.\n1. **Thesis visible in first 100 words?** If not, rewrite the hook.\n2. **Every section has ≥1 specific number, SHA, PR#, filename, or date?** (generic adjectives don't count)\n3. **Zero banned phrases** (see *Banned phrase lexicon* section below — check against that explicit list).\n4. **Counter-evidence is real** — not a strawman like \"some might say it's complex\".\n5. **Sources ≥4 links, ≥1 in-repo, ≥1 external.**\n6. **Title asserts something** (not \"A look at X\" \u002F \"Exploring Y\").\n7. **Word count in 600–900** (hard bounds — trim or expand).\n8. **No placeholder phrases** like \"[TBD]\", \"[link]\", \"[title]\".\n\nIf any item still fails after one rewrite pass, publish with status `REPO_ARTICLE_DEGRADED` and note which items failed in the log — don't hide it.\n\n### Phase 5 — Save, log, notify (repo)\n\n1. Save the article to `output\u002Farticles\u002Frepo-article-${today}.md`. (If a hero image was generated via the Visual add-on, put `![hero](..\u002Fimages\u002Frepo-article-${today}.\u003Cext>)` at the top.)\n2. Append the consolidated log entry (see **Log**) **before** notifying.\n3. Update the `Recent Articles` table in `memory\u002FMEMORY.md` (Date | Title | Topic).\n4. Notify via `.\u002Fnotify`:\n   ```\n   *[Article title]*\n\n   Thesis: [one sentence]\n\n   Read: [link to output\u002Farticles\u002Frepo-article-${today}.md in THIS repo - get the repo name from `git remote get-url origin`, not the watched repo]\n   ```\n\n### Banned phrase lexicon (repo angle)\n\nReject a draft that contains any of these. Match case-insensitively, whole phrase or obvious variant:\n- \"in today's fast-paced world\"\n- \"leveraging\" \u002F \"leverage\" (as a verb meaning \"use\")\n- \"robust\"\n- \"game-changer\" \u002F \"game-changing\"\n- \"under the hood\" (unless the section actually walks through internals)\n- \"taking X to the next level\"\n- \"at the end of the day\"\n- \"diving into\" \u002F \"deep dive\"\n- \"delving into\" \u002F \"delve\"\n- \"comprehensive suite\"\n- \"cutting-edge\"\n- \"seamlessly\" \u002F \"seamless\"\n- \"empowers\" \u002F \"empowering\"\n- \"revolutionize\" \u002F \"revolutionary\"\n- \"unlock\" (metaphorical, e.g. \"unlocks new possibilities\")\n- \"streamline\" (as filler)\n- \"best-in-class\"\n- \"paradigm shift\"\n\nIf a banned phrase is the *most accurate* word in a technical context (e.g. actually describing leverage in a derivatives article), keep it and note the exemption in the log.\n\n### Constraints (repo angle)\n\n- Never publish without a thesis.\n- Never pad to hit word count — 600 honest words beat 900 padded.\n- Never fabricate a SHA, PR number, or quote. If real evidence isn't available, weaken the thesis or skip.\n\n---\n\n## Angle: lens — the project through a surprising lens\n\n\u003C!-- autoresearch: variation B — editorial discipline (research → falsifiable thesis → draft → self-edit with hard gates) -->\n\nWrites articles that explain the project through a **different lens each time** — framed so a reader who's never heard of the project understands why it matters, via something they already care about. NOT a repo progress update (that's the **repo** angle above). `arg` is the lens (e.g. \"unix philosophy\", \"regulation wave\", \"open source funding\"); if empty, auto-select from trending topics + angle rotation.\n\nRead before deciding anything: `memory\u002FMEMORY.md`, the last 7 days of `memory\u002Flogs\u002F`, `memory\u002Fwatched-repos.md`, and `memory\u002Fproject-lens-angles.md` (may not exist on first run — treat absence as empty history).\n\n**Why models fail at this by default:** they slide into feature-listing wrapped in philosophical language, forced parallels with no mechanism, and marketing tone. This angle prevents that with a research → thesis → draft → self-edit pipeline where each phase has hard gates. If the gates can't pass, abort — don't publish a weak article.\n\n### Phase 1 — Context\n\nRead before deciding anything:\n- Last 14 days of `output\u002Farticles\u002Fproject-lens-*.md` and `memory\u002Fproject-lens-angles.md` — know which angle categories and theses are exhausted.\n- 2–3 most recent `output\u002Farticles\u002Frepo-article-*.md` and `output\u002Farticles\u002Fpush-recap-*.md` — know what shipped lately.\n- Repo state: `gh api repos\u002F{owner}\u002F{repo} --jq '{name, description, stargazers_count, forks_count, open_issues_count, updated_at}'`. If unreachable, continue with memory only and log the gap.\n\nIf `memory\u002Fwatched-repos.md` is empty or missing, abort and notify: \"project-lens: no watched repo configured.\"\n\n### Phase 2 — Pick the lens\n\n**If `arg` is set**, use it verbatim. Classify into one of the 8 categories below for logging.\n\n**If `arg` is empty**:\n1. Run 2–3 WebSearch queries on what's being debated right now in tech, crypto, AI, regulation, open source, or philosophy (e.g., `\"AI agents\" autonomy debate last 7 days`, `crypto regulation April 2026`, `open source funding model 2026`).\n2. From results, identify 3 candidate angles with non-obvious connections to the project.\n3. Pick the one that (a) hasn't appeared in the last 14 days **and** (b) has the strongest concrete connection. Record the choice and the rejected candidates with one-line reasons.\n\n**Angle categories (no repeat within 14 days):**\n1. **Current events** — Something happening this week\u002Fmonth.\n2. **Philosophy \u002F big ideas** — Unix philosophy, cathedral vs bazaar, composability, anti-fragility, skin in the game, swarm intelligence, etc.\n3. **Industry comparison** — How a well-known company\u002Fproject solved a similar problem differently.\n4. **User story** — POV of a specific persona (solo dev, DAO, research lab, crypto community) with and without this tool.\n5. **Contrarian take** — Challenge a common assumption; use project as evidence.\n6. **Technical deep-dive for non-technical readers** — One architectural decision, plain language, bigger implications.\n7. **Historical parallel** — Computing \u002F internet \u002F non-tech history with a concrete mechanism (not surface resemblance).\n8. **Ecosystem map** — Where the project sits: adjacent, complementary, competing.\n\n### Phase 3 — Research (gate: collect evidence before drafting)\n\n**External side — required minimums:**\n- 3+ WebSearch queries on the lens topic (different framings, not rewordings)\n- WebFetch on the 2+ most relevant sources\n- ≥3 **distinct domains** across cited sources\n- ≥3 concrete facts extracted: names, numbers, dollar amounts, dated quotes, specific events\n- Recency: ≤30 days old for \"current events\"; ≤180 days for industry comparison \u002F contrarian \u002F ecosystem; ≤5 years for philosophy \u002F historical\n- Log every URL consulted\n\n**Project side — required minimums:**\n- 2+ recent articles in `output\u002Farticles\u002F` read end-to-end\n- `gh api repos\u002F{owner}\u002F{repo}\u002Fcommits --jq '.[0:10] | .[] | {sha: .sha[0:7], msg: (.commit.message|split(\"\\n\")[0])}'` — last 10 commits\n- ≥3 specific project references you plan to use: named features, file paths, commit hashes, architectural choices. **Not** vague claims like \"the project uses AI\" or \"it has good UX.\"\n\n**If you cannot hit these minimums, abandon the angle and re-run Phase 2 with a different category.** Log the abandoned angle and why.\n\n### Phase 4 — Thesis lock (hard gate)\n\nBefore drafting, write **ONE falsifiable claim in ≤30 words** that links the lens to the project. Example:\n> \"Running agents as scheduled GitHub Actions — rather than as persistent servers — trades a few seconds of latency for a property the AI industry barely has: versioned, audit-trailed, publicly forkable autonomy.\"\n\nRules:\n- **Falsifiable**: a reasonable critic could argue the opposite.\n- **Specific**: names concrete things (cron jobs, not \"infrastructure\"; audit-trailed, not \"better\").\n- Not a tautology. Not marketing. Not \"this is cool because X.\"\n\n**If you can't state the thesis in one sentence, the angle isn't working — return to Phase 2.** Do not proceed with a fuzzy thesis.\n\n### Phase 5 — Draft (700–1000 words)\n\nSave to `output\u002Farticles\u002Fproject-lens-${today}.md` with this structure:\n```markdown\n# [Title: leads with the lens, works for a reader who doesn't know the project]\n\n[¶1-2: external hook. Start with the trend\u002Fidea\u002Fevent\u002Fquestion the reader already cares about. Do NOT name the project yet.]\n\n## [Section: establishes the external frame]\n[Build the lens with one or more of your concrete facts - a quote, a number, a specific event.]\n\n## [Section: introduces the project through the frame]\n[Project enters here - but through the lens, not as a feature list. Describe how it embodies, challenges, or extends the idea with specific code\u002Fdesign references.]\n\n## [Section: one non-obvious technical or strategic detail]\n[Where the article earns its existence. Point to something in the code, architecture, or approach a reader wouldn't get from the README.]\n\n## [Section: zoom back out]\n[A concrete forward claim - specific enough to be wrong. Not \"this is exciting.\" Something like \"this suggests X won't happen for 2-3 years because Y\" or \"this is the same mistake [named case] made, and it took [duration] to recover.\"]\n\n---\n*Sources:*\n- [source title](url) - what it was used for\n- ...\n```\n\n**Draft requirements:**\n- Title must work for a reader who doesn't know the project name.\n- ≥3 external citations rendered as inline links.\n- ≥3 specific project references (named features, file paths, commit hashes, named decisions).\n- 700–1000 words (count before submitting).\n\n### Phase 6 — Self-edit (hard gates — all must pass)\n\nGo through this checklist after the first draft. If any gate fails, rewrite the affected section **once**. If the second pass still fails, **abort and log** — do not publish a weak article.\n- [ ] Title does NOT name the project\n- [ ] First 2 paragraphs do NOT name the project\n- [ ] ≥3 external citations with URLs (inline)\n- [ ] ≥3 specific project references (named, not vague)\n- [ ] Falsifiable thesis visible in the article text\n- [ ] 700–1000 words\n- [ ] No banned phrases: *revolutionary*, *groundbreaking*, *game-changing*, *paradigm shift*, *disrupting*, *unlocks*, *empowers*, *the future of X*, *leverage \u002F leveraging*, *at scale*, *democratize* (unless quoting a source that used the word)\n- [ ] Every parallel\u002Fcomparison states a concrete mechanism — not surface resemblance\n- [ ] Closing section makes a specific forward claim (not generic optimism or \"time will tell\")\n- [ ] No feature-list paragraphs (\"the project does X, Y, Z\") — if found, cut and keep ONE element\n\n### Phase 7 — Output (lens)\n\n1. **Save** `output\u002Farticles\u002Fproject-lens-${today}.md`. (If a hero image was generated via the Visual add-on, put `![hero](..\u002Fimages\u002Fproject-lens-${today}.\u003Cext>)` at the top.)\n2. **Append** to `memory\u002Fproject-lens-angles.md` (create if missing):\n   ```markdown\n   ## ${today}\n   - Angle: [category]\n   - Thesis: [one-line falsifiable claim]\n   - Title: [article title]\n   - Sources: [3-5 URLs]\n   ```\n3. **Notify** via `.\u002Fnotify`:\n   ```\n   *New Article: [title]*\n\n   [3-4 sentence summary: the external thing the article connects to, the thesis claim, one specific project detail.]\n\n   Read: [URL to output\u002Farticles\u002Fproject-lens-${today}.md - use `git remote get-url origin` for this repo]\n   ```\n4. **Log** the consolidated entry (see **Log**).\n\n### Anti-patterns (prevention beats self-edit catch)\n\n- **Forced parallels** — every comparison needs a concrete mechanism, not surface resemblance. \"X is like Y because both are new\" fails; \"X is like Y because both decoupled [specific function] from [specific bottleneck]\" passes.\n- **Feature-dump via the lens** — pick ONE architectural decision and interrogate it, don't list what the project does.\n- **Marketing tone** — aim for trade-publication prose, not a company blog.\n- **False novelty** — if you can't point to what's actually new, name what's old that still works and why.\n- **Vague closings** — the final section must make a claim specific enough that a reader could come back in six months and say \"you were wrong\" or \"you were right.\"\n\n### Constraints (lens angle)\n\n- Never publish if Phase 6 self-edit fails twice — abort cleanly.\n- Never reuse an angle category within 14 days (check `memory\u002Fproject-lens-angles.md`).\n- Never invent facts to fill the citation minimum — if research is thin, abandon the angle.\n\n---\n\n## Visual add-on (`--visual`) — Replicate hero image\n\nRuns **only when `visual = true`**, after the article body is written and saved, for **any** angle. Use Replicate's Nano Banana Pro (Gemini 3 Pro Image). It renders **text labels well** — exploit that by writing prompts that ask for labeled diagrams or schematics, not stock-photo metaphors. Set `IMG_BASENAME` to match the article file for the angle: `explainer-${today}` (standard\u002Fexplainer), `article-${today}` (standard\u002Fgeneral), `repo-article-${today}` (repo), or `project-lens-${today}` (lens).\n\n1. **Preflight**: read the \"Keys set for this run\" line in your run notes, never a shell test (those are refused and cost a turn). If `REPLICATE_API_TOKEN` is not on that line, log `IMAGE_SKIPPED reason=no-token` and skip to step 5 (no-image path). Do not attempt any Replicate call. The article must ship without an image in this case.\n\n2. **Craft the prompt**. Aim for technical illustration energy, not marketing. Strong prompt templates:\n   - *Schematic*: \"Technical schematic illustration of \u003Cmechanism>, dark navy background, thin cyan and amber lines, labeled boxes reading '\u003Clabel1>', '\u003Clabel2>', '\u003Clabel3>', arrows showing data flow from \u003CA> to \u003CB> to \u003CC>, blueprint aesthetic, 16:9\"\n   - *Conceptual*: \"Editorial illustration capturing \u003Ccore concept>: \u003Cvisual metaphor with concrete objects>, flat geometric style, restrained palette of two accent colors on near-black background, no human figures, 16:9\"\n   - *Data-flow*: \"Network diagram of \u003Cmechanism>: nodes labeled '\u003CA>', '\u003CB>', '\u003CC>' connected by directional arrows, weights shown as line thickness, monospace labels, technical-paper figure style, 16:9\"\n   Avoid: photorealistic faces, stock-business imagery, \"AI brain\" tropes, gradient slop.\n\n3. **Generate** with fallback enabled from the start (Nano Banana Pro can rate-limit; Seedream 5.0 lite is the fallback). The Replicate call is auth'd, so route it through `.\u002Fsecretcurl` with the `{REPLICATE_API_TOKEN}` placeholder — never a bare `$REPLICATE_API_TOKEN` on the line (the Bash permission layer refuses it):\n   ```bash\n   .\u002Fsecretcurl -s -X POST \\\n     -H \"Authorization: Bearer {REPLICATE_API_TOKEN}\" \\\n     -H \"Content-Type: application\u002Fjson\" \\\n     -H \"Prefer: wait\" \\\n     -d '{\n       \"input\": {\n         \"prompt\": \"YOUR_DETAILED_PROMPT_HERE\",\n         \"aspect_ratio\": \"16:9\",\n         \"number_of_images\": 1,\n         \"safety_tolerance\": 5,\n         \"allow_fallback_model\": true\n       }\n     }' \\\n     \"https:\u002F\u002Fapi.replicate.com\u002Fv1\u002Fmodels\u002Fgoogle\u002Fnano-banana-pro\u002Fpredictions\"\n   ```\n   `Prefer: wait` usually returns the image inline in `.output`. If `.output` is empty, the prediction is still running - poll `.urls.get` for up to ~60s (`.\u002Fsecretcurl -s -H \"Authorization: Bearer {REPLICATE_API_TOKEN}\" \"\u003CPRED_URL>\"`, the `.urls.get` value written in literally), stopping when `.status` is `succeeded` (read `.output`) or `failed`\u002F`canceled` (no-image path, step 5).\n\n4. **Persist locally** — Replicate CDN URLs expire. Download and commit (the CDN URL carries no secret, so plain `curl` is fine):\n   ```bash\n   mkdir -p output\u002Fimages\n   ```\n   `IMAGE_URL` = the URL extracted from `response.output`. `EXT` = the last `.jpg`, `.jpeg`, `.png` or `.webp` in IMAGE_URL, or `.jpg` if it has none (work it out yourself). `LOCAL_PATH` = `output\u002Fimages\u002F\u003CIMG_BASENAME>\u003CEXT>`. Then download with every value written in literally:\n   ```bash\n   curl -sL \"\u003CIMAGE_URL>\" -o \"\u003CLOCAL_PATH>\"\n   ```\n   If the curl fails, retry via WebFetch or skip (no-image path, step 5).\n\n5. **No-image path** (token missing, API down, rate-limited, or download failed): log `IMAGE_SKIPPED reason=\u003Cconcrete reason>` and proceed with the article. Add a one-line note at the top of the article: `\u003C!-- hero image skipped: \u003Creason> -->`. The text must stand on its own. Never fail the whole skill because of an image problem.\n\nOnce the image is saved, add the hero-image line to the top of the article file (`![hero](..\u002Fimages\u002F${IMG_BASENAME}.\u003Cext>)`) and include the original Replicate URL in that angle's notification if the angle's notify format has an image slot.\n\n---\n\n## Log\n\nAppend **one** entry under a single `### article` heading in `memory\u002Flogs\u002F${today}.md`, as bullet points. Start with a discriminator line naming the branch\u002Fmode that ran, then the branch-specific fields:\n\n```\n### article\n- Branch: standard | repo | lens   (+visual if --visual ran)\n```\n\n**Standard branch fields:**\n```\n- Mode: general-article | technical-explainer\n- Topic: [topic]\n- Title: [title]\n- Key idea: [one-sentence claim]   (technical-explainer only)\n- Image: generated | fallback-model | skipped (\u003Creason>) | n\u002Fa\n- Image prompt: [prompt used, or \"n\u002Fa\"]\n- Primary source: [URL]            (technical-explainer only)\n- File: output\u002Farticles\u002F${today}.md | output\u002Farticles\u002Fexplainer-${today}.md\n- Notification sent: yes | no\n```\n\n**Repo branch fields:**\n```\n- Repo: owner\u002Frepo\n- Thesis: [verbatim]\n- Angle: [arg or auto-selected]\n- Word count: N\n- Self-edit checklist: X\u002F8 passing\n- Image: generated | fallback-model | skipped (\u003Creason>) | n\u002Fa\n- Status: REPO_ARTICLE_OK | REPO_ARTICLE_DEGRADED | REPO_ARTICLE_SKIPPED\n```\n\n**Lens branch fields:**\n```\n- Angle: [category]\n- Thesis: [one-line]\n- External sources: [count] across [N] distinct domains\n- Project references: [count]\n- Self-edit gates: all passed | failed at [gate name] → rewrite → [passed | aborted]\n- Image: generated | fallback-model | skipped (\u003Creason>) | n\u002Fa\n- Status: published | aborted\n- Notification: sent | skipped\n```\n\nLog **always — even on partial failure** (e.g. IMAGE_SKIPPED, REPO_ARTICLE_SKIPPED, aborted lens).\n\n## Network note\n\nThere is no network sandbox — `curl` works. For a flaky public GET, fall back to **WebFetch** on the same URL. For an auth'd API, call `.\u002Fsecretcurl` with a `{ENV_NAME}` placeholder (the key is injected via `requires:`) — never a bare `$SECRET` on the line. `gh api` handles GitHub auth internally — prefer it over raw curl for repo metadata.\n\n**Shell shape.** Each Bash call is one allowed command (pipes are fine). The run refuses a call that sets a variable (`EXT=$(...)`, `IMAGE_URL=...`), contains a `$(...)` substitution, uses a `[ ... ]` test, or joins commands with `;`, `&&` or `||` (including an `|| echo` fallback), and each refusal costs a model turn. So every command in this skill is its own call: a `date` lookup prints a value you write into later commands literally (`\u003CSINCE>`, `\u003CIMAGE_URL>`, `\u003CEXT>`), the token preflight reads the run notes' \"Keys set for this run\" line instead of a shell test, and a fallback is a prose step you take when the call fails.\n\nThe Replicate call runs **in-run** via `.\u002Fsecretcurl` (see the Visual add-on). If it fails, times out, or the download fails, go straight to the no-image path (step 5) — the article ships text-only. There is no deferred fallback; the image is best-effort and never blocks the article.\n\n## Environment Variables\n- `REPLICATE_API_TOKEN` — Replicate API key, used only by the `--visual` add-on. Optional: article text ships without it via the no-image path.\n\nWrite complete, publication-ready content. No placeholders.\n",{"name":5,"metadata":57,"description":7},{"var":58,"tags":59,"title":61,"category":32,"requires":62},"",[60,42],"content","Article",[63],"REPLICATE_API_TOKEN?","main","504a21d53c0c9b485f6e90c43db83000825e1d3c","skills\u002Farticle",308,"https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon","https:\u002F\u002Fwww.aeon.fun\u002F","https:\u002F\u002Fgithub.com\u002Faeonfun","The most autonomous agent framework.","2026-11-15 03:40:18",{"source":8,"skillName":5,"commands":74,"prompt":84,"zipAvailable":85},[75,78,81],{"tool":76,"command":77},"npx","npx skills add aeonfun\u002Faeon --skill article",{"tool":79,"command":80},"bunx","bunx skills add aeonfun\u002Faeon --skill article",{"tool":82,"command":83},"pnpm","pnpm dlx skills add aeonfun\u002Faeon --skill article","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon\u002Fblob\u002Fmain\u002Fskills\u002Farticle\u002FSKILL.md",true,[],[88,95,101,107,113,119],{"skillNo":89,"name":90,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":94},"01M7DSM3QWBJ44FKRMDDRAGNET","Warn Mixed Script","https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Fskill-scan-fixtures\u002Fwarn-mixed-script\u002FSKILL.md",[],[],"2026-11-15 14:19:17",{"skillNo":96,"name":97,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":98,"categories":99,"keywords":100,"publishedAt":94},"01M7DSM0496DRYYW5A4NAAHZDX","Warn Hidden CSS","https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Fskill-scan-fixtures\u002Fwarn-hidden-css\u002FSKILL.md",[],[],{"skillNo":102,"name":103,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":94},"01M7DSKWCGDVYZSDJM31YHS2JB","Malicious Injection Uppercase","https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Fskill-scan-fixtures\u002Fmalicious-injection-uppercase\u002FSKILL.md",[],[],{"skillNo":108,"name":109,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":110,"categories":111,"keywords":112,"publishedAt":94},"01M7DSKRV2RRA1V0HK13KYWJXD","Malicious Injection Homoglyph","https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Fskill-scan-fixtures\u002Fmalicious-injection-homoglyph\u002FSKILL.md",[],[],{"skillNo":114,"name":115,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":116,"categories":117,"keywords":118,"publishedAt":94},"01M7DSKMXJKY0167YHVCKF7B49","Malicious Injection Disregard Prior","https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Fskill-scan-fixtures\u002Fmalicious-injection-disregard-prior\u002FSKILL.md",[],[],{"skillNo":120,"name":121,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":122,"categories":123,"keywords":124,"publishedAt":94},"01M7DSKGFQ5V2VNSA4K97CVFMB","Benign Greek Math","https:\u002F\u002Fgithub.com\u002Faeonfun\u002Faeon\u002Fblob\u002Fmain\u002Fscripts\u002Ftests\u002Fskill-scan-fixtures\u002Fbenign-greek-math\u002FSKILL.md",[],[]]