Update to Arcads

{
  "name": "spy-competitor-ads",
  "description": "Find and download competitor ads from the Meta Ad Library (video, image, or both). Invoke with /arcads:spy-competitor-ads or via another skill.",
  "included_files": [],
  "skill_md_contents": "---\nname: spy-competitor-ads\ndescription: Find and download competitor ads from the Meta Ad Library (video, image, or both). Invoke with /arcads:spy-competitor-ads or via another skill.\n---\n\n# Spy Competitor Ads — Find & Download\n\nYour only job: **find competitor ads and download them.** Nothing else. Execute the whole pipeline silently and the user's first output is the downloaded creatives themselves.\n\n**Media type — pick exactly ONE of three modes. NO default.**\n\n| Mode | Explicit trigger phrases | `media_type` param | Extractor | Downloads |\n|---|---|---|---|---|\n| **VIDEO** | \"video ads\", \"competitor videos\", \"video creatives\", \"their videos\" | `video` | `<video>` only | `.mp4` |\n| **IMAGE** | \"static ads\", \"image ads\", \"photo ads\", \"still creatives\", \"posters\", \"image creatives\" | `image` | `<img>` only (min 200px) | `.jpg` / `.png` |\n| **BOTH** | \"both\", \"video and image\", \"videos and statics\", \"all formats\", \"everything\", \"any media\", \"full swipe file\" | `all` | `<video>` + `<img>` in one pass | `.mp4` + `.jpg` / `.png` |\n\nIf — and only if — the user's request **does not contain an explicit trigger for one of the three modes**, ask with `AskUserQuestion` before doing anything else (see Step 1b). Do NOT silently default to video. The chosen mode then flows through the whole skill: URL → in-page extractor → file extensions → delivery list. Pick once and stick with it for every competitor in the run.\n\n---\n\n## Golden rules\n\n1. **Silent execution. No narration, no logging.** Never say what you are doing or have done — no \"Searching…\", \"Found 12 ads\", \"Downloading…\", \"Moving files…\". No status lines, no step commentary. The user sees only the final downloaded creatives.\n2. **Up to three forms, and only these three.** You may ask: (a) which media mode to run (Step 1b, if not explicit in the request), (b) which auto-found competitors to keep (Step 1c, when you had to discover them yourself), and (c) which delivered creatives to clone (Step 5, only on standalone runs). All use `AskUserQuestion`. Beyond these three, no other questions, confirmations, or \"proceed?\" check-ins. Between forms, go fully autonomous.\n3. **No analysis.** Do NOT analyze the creatives, describe hooks, summarize messaging, rank by \"why it's winning\", or write a competitive brief. Do not profile the competitors. Just find and download.\n4. **Auto-find competitors when not given — then confirm.** If the user names competitors, use them as-is, no confirmation. If not, find them yourself AND surface them with a multi-select form so the user can drop any that don't fit (Step 1c). Never silently scrape against a list the user never saw.\n5. **No technical leakage.** Never mention CDN URLs, asset IDs, MCP tool names, scraping mechanics, or file-move steps.\n6. **One mode per run** (VIDEO, IMAGE, or BOTH — see the mode table above). Skip carousels in every mode.\n7. **Browser MCP is required.** The Meta Ad Library is JavaScript-rendered. If no browser automation MCP is connected, that is the one thing you stop and report (Step 2).\n\n---\n\n## Step 1 — Resolve the brand, mode, and competitors\n\nYou run up to three short setup interactions here, **in this exact order**, and only the ones whose answer isn't already in the request. Once these are settled, go autonomous.\n\n### 1a — Brand\n\n- **Brand named or obvious from context?** → use it.\n- **Not inferable?** → ask exactly one short question: *\"What's your brand or product?\"* Nothing else.\n\n### 1b — Media mode (NO default — ask if missing)\n\nRead the user's request. If it contains an explicit phrase from the mode-trigger table (VIDEO / IMAGE / BOTH), use that and skip this step.\n\nIf the request is silent on media type, ask with `AskUserQuestion`:\n\n> **Question**: \"Which kind of competitor ads should I grab?\"\n> **Header**: \"Media type\"\n> **Options**:\n> - **Video ads** — \"Playable video creatives only (.mp4)\"\n> - **Static / image ads** — \"Static image creatives only (.jpg / .png)\"\n> - **Both video and image** — \"Everything they're running, in one pass (Recommended)\"\n\nWait for the answer before doing anything else. **Do not default to video.** Do not scrape without an explicit choice.\n\n### 1c — Competitor shortlist\n\n- **Competitors named by the user?** → use them as-is. Skip this confirmation entirely.\n- **Not named?** → find 3–5 direct competitors yourself with a quick `WebSearch`/`WebFetch` (brands selling a similar product to a similar audience that plausibly run paid social). Then confirm with `AskUserQuestion`:\n\n  > **Question**: \"Found these competitors for [user's brand]. Pick the ones to spy on — keep the relevant ones, drop the rest.\"\n  > **Header**: \"Competitors\"\n  > **Multiple**: `true`\n  > **Options**: one per discovered competitor, with a one-line label like \"BrandX — meal kits\", \"BrandY — recipe app\", etc. Each `description` explains why it was matched.\n\n  The form's `custom` option lets the user type any competitor you missed. Run the scrape against the final selection (kept + added). If the user selects zero, ask once for a manual list, then stop.\n\n  Never silently scrape against an auto-found list the user never saw.\n\n### Count\n\nDefault count: top **5** ads pooled across the final competitor selection. If the user specified a number (\"2 ads\", \"one each\"), honor it exactly.\n\n---\n\n## Step 2 — Browser MCP check\n\nConfirm a browser automation MCP is connected (e.g. `mcp__Claude_in_Chrome__*`, Playwright, Chrome DevTools, Puppeteer). Verify with `list_connected_browsers` or equivalent.\n\n- **Connected** → proceed silently.\n- **Not connected** → stop and say only:\n  > \"I need a browser automation plugin (like the Claude-in-Chrome extension or Playwright) connected to read the Meta Ad Library. Connect one and I'll grab the ads.\"\n\nDo not try to work around this with `WebFetch`.\n\n---\n\n## Step 3 — Scrape + download in one pass per competitor\n\nRun competitors **sequentially** (parallel sessions trigger bot detection).\n\nFor each competitor, build the URL — pick `media_type` based on the chosen mode:\n\n**VIDEO mode (default):**\n```\nhttps://www.facebook.com/ads/library/?active_status=active&ad_type=all&country=ALL&is_targeted_country=false&media_type=video&search_type=keyword_unordered&sort_data[direction]=desc&sort_data[mode]=total_impressions&q=<COMPETITOR_URL_ENCODED>\n```\n\n**IMAGE mode:**\n```\nhttps://www.facebook.com/ads/library/?active_status=active&ad_type=all&country=ALL&is_targeted_country=false&media_type=image_and_meme&search_type=keyword_unordered&sort_data[direction]=desc&sort_data[mode]=total_impressions&q=<COMPETITOR_URL_ENCODED>\n```\n\n**BOTH mode (image + video):**\n```\nhttps://www.facebook.com/ads/library/?active_status=active&ad_type=all&country=ALL&is_targeted_country=false&media_type=all&search_type=keyword_unordered&sort_data[direction]=desc&sort_data[mode]=total_impressions&q=<COMPETITOR_URL_ENCODED>\n```\n\nUse the user's primary market for `country=` if they mentioned one, else `ALL`.\n\nThen, for speed, do everything in as few calls as possible — **navigate, wait briefly, then one JavaScript call** that scrolls, extracts, filters, and downloads:\n\n### Critical technical facts (learned the hard way)\n\n- **`curl` does NOT work.** The browser extension masks Meta CDN media URLs (they carry cookie/query-string/JWT tokens), so the raw `src` is never returned to you and an external `curl` has no valid URL. **Download in-page instead**: `fetch(src)` → `blob()` → temporary `<a download>` → `click()`. This saves to the browser's Downloads folder. You never need to see the URL.\n- **Keyword pollution is common.** A search for a brand often returns an unrelated company with the same name (e.g. \"Creatify\" the AI tool vs. \"Creatify.mx\" a sticker shop). Inspect each card's advertiser name / domain / copy and keep only cards that match the real competitor. Drop the rest.\n- **Impressions are hidden** for commercial (non-political) ads. Don't rank by impressions. Instead prefer the **most-recurring creative** (many near-identical live copies = highest spend = proven winner), then most recent. Pick the top N by that heuristic.\n- **Selector + extension depend on the mode.** VIDEO mode uses `<video>` + `.mp4`. IMAGE mode uses the card's main `<img>` + `.jpg`/`.png`. BOTH mode collects `<video>` and `<img>` in the same pass (with the same min-size filter for images). In every mode, filter out tiny avatars/icons (≤ ~200px on either side) so you only keep real ad creatives.\n\n### One-shot extract + download script — VIDEO mode\n\n```js\n(async () => {\n  // 1. trigger lazy-load\n  window.scrollTo(0, document.body.scrollHeight);\n  await new Promise(r => setTimeout(r, 2500));\n  window.scrollTo(0, document.body.scrollHeight);\n  await new Promise(r => setTimeout(r, 2000));\n\n  // 2. collect videos + their card text\n  const vids = Array.from(document.querySelectorAll('video'))\n    .filter(v => (v.src || v.currentSrc || '').startsWith('http'));\n  const cardText = (v) => {\n    let el = v;\n    for (let i = 0; i < 12 && el; i++) {\n      el = el.parentElement;\n      if (el && el.innerText && el.innerText.length > 100 && el.innerText.length < 2000) return el.innerText;\n    }\n    return '';\n  };\n\n  // 3. keep only cards matching the REAL competitor (edit the regex per brand),\n  //    drop same-name pollution, and de-dupe so recurring creatives count once.\n  const BRAND = /creatify\\.?ai|@creatify/i;        // <-- set per competitor\n  const kept = [];\n  const seen = new Set();\n  for (const v of vids) {\n    const t = cardText(v);\n    if (!BRAND.test(t)) continue;\n    const key = t.slice(0, 80);\n    kept.push({ v, t, dup: seen.has(key) });\n    seen.add(key);\n  }\n\n  // 4. download up to N (recurring creatives appear first → already spend-weighted)\n  const N = 5;                                      // <-- set to requested count\n  const picks = kept.slice(0, N);\n  const results = [];\n  for (let i = 0; i < picks.length; i++) {\n    const url = picks[i].v.src || picks[i].v.currentSrc;\n    try {\n      const b = await (await fetch(url)).blob();\n      const a = document.createElement('a');\n      a.href = URL.createObjectURL(b);\n      a.download = `spy-ad-${i + 1}-COMPETITOR.mp4`;   // <-- set competitor slug\n      document.body.appendChild(a); a.click(); a.remove();\n      results.push({ i: i + 1, bytes: b.size });\n    } catch (e) { results.push({ i: i + 1, error: String(e) }); }\n  }\n  return results;\n})()\n```\n\n### One-shot extract + download script — STATIC / IMAGE mode\n\n```js\n(async () => {\n  // 1. trigger lazy-load\n  window.scrollTo(0, document.body.scrollHeight);\n  await new Promise(r => setTimeout(r, 2500));\n  window.scrollTo(0, document.body.scrollHeight);\n  await new Promise(r => setTimeout(r, 2000));\n\n  // 2. collect images + their card text. Drop avatars/icons by minimum size.\n  const MIN = 200;                                  // px — anything smaller is likely an avatar/icon\n  const imgs = Array.from(document.querySelectorAll('img'))\n    .filter(i => (i.currentSrc || i.src || '').startsWith('http'))\n    .filter(i => (i.naturalWidth || i.width) >= MIN && (i.naturalHeight || i.height) >= MIN);\n  const cardText = (n) => {\n    let el = n;\n    for (let i = 0; i < 12 && el; i++) {\n      el = el.parentElement;\n      if (el && el.innerText && el.innerText.length > 100 && el.innerText.length < 2000) return el.innerText;\n    }\n    return '';\n  };\n\n  // 3. keep only cards matching the REAL competitor (edit the regex per brand),\n  //    drop same-name pollution, and de-dupe so recurring creatives count once.\n  const BRAND = /creatify\\.?ai|@creatify/i;        // <-- set per competitor\n  const kept = [];\n  const seen = new Set();\n  for (const img of imgs) {\n    const t = cardText(img);\n    if (!BRAND.test(t)) continue;\n    const key = t.slice(0, 80);\n    if (seen.has(key)) continue;                    // de-dupe identical creatives\n    kept.push({ img, t });\n    seen.add(key);\n  }\n\n  // 4. download up to N as JPGs (recurring creatives appear first → already spend-weighted)\n  const N = 5;                                      // <-- set to requested count\n  const picks = kept.slice(0, N);\n  const results = [];\n  for (let i = 0; i < picks.length; i++) {\n    const url = picks[i].img.currentSrc || picks[i].img.src;\n    try {\n      const b = await (await fetch(url)).blob();\n      const ext = (b.type && b.type.includes('png')) ? 'png' : 'jpg';\n      const a = document.createElement('a');\n      a.href = URL.createObjectURL(b);\n      a.download = `spy-ad-${i + 1}-COMPETITOR.${ext}`;  // <-- set competitor slug\n      document.body.appendChild(a); a.click(); a.remove();\n      results.push({ i: i + 1, bytes: b.size, ext });\n    } catch (e) { results.push({ i: i + 1, error: String(e) }); }\n  }\n  return results;\n})()\n```\n\n### One-shot extract + download script — BOTH mode (image + video)\n\nCollects videos and images in a single pass off the same `media_type=all` page. Deduplication is per-card (one creative per ad card, regardless of whether it's a video or an image), so a single card with both a poster image and a playable video counts once and prefers the video.\n\n```js\n(async () => {\n  // 1. trigger lazy-load\n  window.scrollTo(0, document.body.scrollHeight);\n  await new Promise(r => setTimeout(r, 2500));\n  window.scrollTo(0, document.body.scrollHeight);\n  await new Promise(r => setTimeout(r, 2000));\n\n  const MIN = 200;                                  // px — minimum image size to count as a real creative\n\n  // 2. helper: walk up to find the ad card and its text, return both\n  const cardOf = (node) => {\n    let el = node;\n    for (let i = 0; i < 12 && el; i++) {\n      el = el.parentElement;\n      if (el && el.innerText && el.innerText.length > 100 && el.innerText.length < 2000) {\n        return { card: el, text: el.innerText };\n      }\n    }\n    return { card: null, text: '' };\n  };\n\n  // 3. collect candidates: every video, plus every large image\n  const candidates = [];\n  for (const v of document.querySelectorAll('video')) {\n    const url = v.src || v.currentSrc || '';\n    if (!url.startsWith('http')) continue;\n    candidates.push({ kind: 'video', node: v, url });\n  }\n  for (const img of document.querySelectorAll('img')) {\n    const url = img.currentSrc || img.src || '';\n    if (!url.startsWith('http')) continue;\n    if ((img.naturalWidth || img.width) < MIN || (img.naturalHeight || img.height) < MIN) continue;\n    candidates.push({ kind: 'image', node: img, url });\n  }\n\n  // 4. attach card text + de-dupe per card (video wins over image when both exist on the same card)\n  const BRAND = /creatify\\.?ai|@creatify/i;          // <-- set per competitor\n  const byCard = new Map();                          // card element → chosen candidate\n  for (const c of candidates) {\n    const { card, text } = cardOf(c.node);\n    if (!card) continue;\n    if (!BRAND.test(text)) continue;\n    const prev = byCard.get(card);\n    // prefer video over image when both exist on the same card\n    if (!prev || (prev.kind === 'image' && c.kind === 'video')) {\n      byCard.set(card, { ...c, text });\n    }\n  }\n\n  // 5. de-dupe near-identical recurring creatives by card text prefix\n  const seen = new Set();\n  const kept = [];\n  for (const c of byCard.values()) {\n    const key = c.text.slice(0, 80);\n    if (seen.has(key)) continue;\n    seen.add(key);\n    kept.push(c);\n  }\n\n  // 6. download up to N (recurring creatives appear first → already spend-weighted)\n  const N = 5;                                       // <-- set to requested count\n  const picks = kept.slice(0, N);\n  const results = [];\n  for (let i = 0; i < picks.length; i++) {\n    const p = picks[i];\n    try {\n      const b = await (await fetch(p.url)).blob();\n      let ext;\n      if (p.kind === 'video') {\n        ext = 'mp4';\n      } else {\n        ext = (b.type && b.type.includes('png')) ? 'png' : 'jpg';\n      }\n      const a = document.createElement('a');\n      a.href = URL.createObjectURL(b);\n      a.download = `spy-ad-${i + 1}-COMPETITOR.${ext}`;  // <-- set competitor slug\n      document.body.appendChild(a); a.click(); a.remove();\n      results.push({ i: i + 1, kind: p.kind, bytes: b.size, ext });\n    } catch (e) { results.push({ i: i + 1, error: String(e) }); }\n  }\n  return results;\n})()\n```\n\nIf a fetch fails (expired/geo-blocked), it's skipped automatically — just move to the next card. No mention to the user.\n\n---\n\n## Step 4 — Collect files and deliver\n\nAfter downloading, move the files out of the browser's Downloads folder to `/tmp/` with the Bash tool, in one command. Pick the glob that matches the mode you ran:\n\n**VIDEO mode:**\n```bash\ncd ~/Downloads && mv -f spy-ad-*.mp4 /tmp/ && ls -la /tmp/spy-ad-*.mp4\n```\n\n**IMAGE mode:**\n```bash\ncd ~/Downloads && mv -f spy-ad-*.jpg spy-ad-*.png /tmp/ 2>/dev/null; ls -la /tmp/spy-ad-*.{jpg,png} 2>/dev/null\n```\n\n**BOTH mode (image + video):**\n```bash\ncd ~/Downloads && mv -f spy-ad-*.mp4 spy-ad-*.jpg spy-ad-*.png /tmp/ 2>/dev/null; ls -la /tmp/spy-ad-*.{mp4,jpg,png} 2>/dev/null\n```\n\nThen deliver — **only the files, no analysis, no commentary, no brief**. Present each creative inline if a preview tool is available; otherwise list them as clickable file links:\n\n> - [spy-ad-1-creatify-ai.mp4](/tmp/spy-ad-1-creatify-ai.mp4)\n> - [spy-ad-2-captions.mp4](/tmp/spy-ad-2-captions.mp4)\n\n…in IMAGE mode:\n\n> - [spy-ad-1-creatify-ai.jpg](/tmp/spy-ad-1-creatify-ai.jpg)\n> - [spy-ad-2-creatify-ai.jpg](/tmp/spy-ad-2-creatify-ai.jpg)\n\n…in BOTH mode (videos and images interleaved by rank):\n\n> - [spy-ad-1-creatify-ai.mp4](/tmp/spy-ad-1-creatify-ai.mp4)\n> - [spy-ad-2-creatify-ai.jpg](/tmp/spy-ad-2-creatify-ai.jpg)\n> - [spy-ad-3-creatify-ai.mp4](/tmp/spy-ad-3-creatify-ai.mp4)\n\nDo not append observations, patterns, recommendations, or written analysis. The only thing that may follow this delivery is the Step 5 \"clone which?\" form — and only when this skill ran standalone.\n\n---\n\n## Step 5 — Offer to clone (standalone runs only)\n\nThis step **only fires when arcads:spy-competitor-ads was invoked directly by the user as the end goal**, not when it was triggered as a sub-step by another skill. Detect the situation before running it:\n\n**Skip Step 5 (do NOT show the form) when any of these is true:**\n- The current invocation was triggered by another skill — most commonly `arcads:clone-hook` (Step 1B auto-source) or `arcads:clone-static-ad` (Step 1B auto-source). Those skills will consume the downloaded files themselves; offering to clone again would loop.\n- The original user request was explicitly research-only — \"just download competitor ads\", \"build me a swipe file\", \"save these for me\", \"I want to look at them\".\n- Zero creatives were delivered in Step 4.\n\n**Run Step 5 (show the form) otherwise**, including when the user's request was ambiguous-but-actionable like \"spy on competitors and let me work from there\", \"find competitor ads I can use as inspiration\", or a bare \"find competitor ads from [brand]\". When in doubt, run it — declining is one click for the user.\n\n### The form\n\nSurface every delivered creative as a separate option in a single `AskUserQuestion` call with `multiple: true`:\n\n> **Question**: \"Want me to clone any of these for your brand? Pick the ones to recreate — I'll rebuild each for you.\"\n> **Header**: \"Clone which?\"\n> **Multiple**: `true`\n> **Options**: one per delivered file, in the same order they were listed in Step 4. Each option's `label` is the short filename without the leading slug (e.g. \"Ad #1 — BrandX (video)\" or \"Ad #3 — BrandY (image)\"). Each option's `description` is one short cue from the card text if you have it (e.g. \"Talking-head selfie, 22s\"), or just the file type if you don't.\n\nThe `custom` default option (\"Type your own answer\") lets the user say \"all of them\", \"none — I'm done\", or a custom note.\n\n### Routing each selection\n\nWalk the user's selection in the order they were picked. For each chosen file:\n\n1. **Video file (`.mp4`)** → load and run the `arcads:clone-hook` skill, passing the local `/tmp/spy-ad-*.mp4` path as the source video (skipping its Step 1B auto-source, since you already have a video). `arcads:clone-hook` will analyze the hook, then (after its own confirmation prompt) clone it for the user's brand.\n2. **Image file (`.jpg` / `.png`)** → load and run the `arcads:clone-static-ad` skill, passing the local `/tmp/spy-ad-*.{jpg,png}` path as the reference static ad (skipping its Step 1B auto-source, since you already have an image). `arcads:clone-static-ad` will analyze the layout and clone it for the user's brand.\n\nRun the chained skills **sequentially** (not in parallel) — each one needs the user's attention for brand/asset questions, and parallel runs would collide. After each chained skill finishes, move to the next selected file.\n\nIf the user picks just one creative, hand off immediately without re-confirming. If they pick \"none\", stop cleanly — no chaining, no further commentary. If they pick \"all of them\", chain through every delivered file in order.\n\n### Carry brand context forward\n\nThe downstream skills (`arcads:clone-hook`, `arcads:clone-static-ad`) will ask for the user's brand, product, and assets. If the user already gave that information during Step 1a or in the original request, pass it forward — don't make them re-answer.\n\n---\n\n## Edge cases\n\n- **No ads of the requested mode found for a competitor** → skip silently, continue with the rest. If *no* competitor yields any creative in the chosen mode, say so briefly and stop. Do **not** silently fall back to a different mode — the user picked VIDEO, IMAGE, or BOTH for a reason.\n- **BOTH mode but only one media type returns** (e.g. competitor runs only videos): deliver what you got and say briefly \"BrandX is only running videos right now — no statics in their library.\" Do not pad with the missing format.\n- **Only same-name/unrelated ads found** → treat as \"no ads found\" for that competitor; skip silently.\n- **Browser not connected** → the only blocking case; see Step 2.\n- **User dismisses the mode form** (Step 1b) or doesn't pick a mode → stop and say \"I need to know which kind of ads to grab — video, image, or both.\" Do not guess.\n- **User deselects every competitor** in the Step 1c multi-select → ask once for a manual list (\"Which competitors should I look at instead?\"). If still empty, stop.\n- **User picks \"none\" in the Step 5 clone form** → stop cleanly. No commentary, no follow-up.\n- **User picks one creative in Step 5** → hand off to the matching cloner skill (`arcads:clone-hook` for video, `arcads:clone-static-ad` for image) without an extra confirmation.\n- **Step 5 chain — one of the cloner skills fails or is cancelled by the user** → stop the chain (do not run the remaining selections silently). Surface what happened in one line and let the user restart the chain if they want.\n- **Invoked from inside another skill** (e.g. `arcads:clone-hook` auto-source) → skip Step 5 entirely. The calling skill owns the next step.\n\n---\n\n## Quick reference\n\n| Tool | Where |\n|---|---|\n| `AskUserQuestion` | Step 1b — pick media mode (when not explicit). Step 1c — multi-select competitor shortlist (when auto-discovered). Step 5 — multi-select \"clone which?\" (standalone runs only). |\n| `WebSearch` / `WebFetch` | Step 1c — auto-find competitors (only if not named) |\n| Browser MCP (`mcp__Claude_in_Chrome__*` / Playwright) | Steps 2–3 — open Ad Library, run extract+download JS |\n| `javascript_tool` (in-page `fetch`→blob→download) | Step 3 — the ONLY reliable download path; curl does not work. Pick the extractor that matches the mode: `<video>` for VIDEO, `<img>` for IMAGE, combined for BOTH. |\n| `Bash` (`mv`) | Step 4 — move files from Downloads to `/tmp/` (`.mp4` for VIDEO, `.jpg`/`.png` for IMAGE, both for BOTH) |\n| `arcads:clone-hook` skill | Step 5 — chained per selected video, passing the local `/tmp/spy-ad-*.mp4` path |\n| `arcads:clone-static-ad` skill | Step 5 — chained per selected image, passing the local `/tmp/spy-ad-*.{jpg,png}` path |\n"
}