Papercut Script

Before you start: read this whole install prompt, tell me in a few lines what this skill does and what installing it involves, and wait for my go-ahead before you do anything else. Prerequisites: none for the install itself: no account, no API key, no paid service, and nothing else to install first. This is a text-only skill: 3 Markdown files, no code, no scripts, and nothing to run. For using it afterwards, read this first, because the folder is NOT the whole system on its own: a full run needs (1) a source link from me to research, (2) the BRIEF at the top of episodes/NN-slug/research/facts.md and the HOOK block at the top of episodes/NN-slug/papercut/STORYBOARD.md, which come from the earlier papercut-strategy and papercut-hook stages (sibling skills, not included in this folder — if either is missing, SKILL.md says to run those first), (3) a SERIES.md at my series root (audience level and region; an example sits in the repository at examples/paper-trail/SERIES.md, outside this folder), (4) web fetch or search plus a browser tool for screenshots, for the research step (the sends this causes are disclosed in point 2; without those tools, I provide the screenshots and facts myself), and (5) only optionally, for choosing sets, the backdrop sheet library/papercut/public/bg/backdrops/_sheet.jpg from the full repository library, resolved via $PAPERCUT_LIB — that sheet is outside this folder, is not fetched by this install, and opening it costs about 0.7-1.5k tokens per image, per SKILL.md. Install "papercut-script" for me. It turns a link into a papercut reel script, all written together because each part feeds the others: a researched facts list with a source for every fact, a 75-95 word script for 22-30 seconds (about 3.5-4 words a second) on a fixed 10-beat spine (hook, proof or tension, reframe, name reveal, how it works, escalation with a stack of exactly three, peak, honest catch, payoff, CTA), a banner fact track of 10-18 banners (3-7 words, ALL CAPS, one fact each, never repeating the voiceover), a staging table with one set per banner, and a humour pass (one gag per banner, from the comedy and format banks). It writes research/facts.md, narration/script.json and the staging table in papercut/STORYBOARD.md, and a script-only run stops there. Repository: https://github.com/0xS4N4TG/papercut-explainer (the skill is the whole skills/papercut-script/ folder at the pinned commit: SKILL.md plus its references/ folder, 3 files) Pinned version: install the code at commit ff8223abc8136cc09e8d7e9b62ecb5d080c389e1 exactly. Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Clone https://github.com/0xS4N4TG/papercut-explainer into a temporary folder, check out commit ff8223abc8136cc09e8d7e9b62ecb5d080c389e1, and copy the skills/papercut-script/ folder from that checkout. Then compare every file in the folder with the fingerprints recorded for this listing on 2026-10-19: [FINGERPRINTS] - SKILL.md: SHA-256 f024173c96e31fe33ac833438aa35e9b5571ebede32e2e1ca0d609ff1b86c8e0 - references/comedy.md: SHA-256 bcf8db5e275cb718f3fe8226bcc71788a89b898adcbd5d5cbd4b4e3f17cfa77f - references/storytelling.md: SHA-256 422ae3d0047d7c85f88b0df472503aaa77c2b9d637144a0e0e46060e010974c7 If a file does not match, it changed since the listing was prepared: do not install it yet, tell me what you found and let me decide. 2. Read SKILL.md in full and read both reference files in references/ (storytelling.md, comedy.md), and tell me in plain words: what the skill does when used, which services it connects to, whether it reads or sends any personal information or keys, and anything that seems unusual. Stop on any red flag and tell me before going further. Confirm these points from the files: - The skill folder holds 3 files only: SKILL.md (6,823 bytes) and two Markdown references in references/ (storytelling.md 30,192 bytes, comedy.md 4,942 bytes). There is no code, no script, no binary and no URL anywhere in the folder: nothing in it can run, call an API, or phone home by itself. All network activity at use time is done by you, the agent, using your own web fetch, search and browser tools, as disclosed below. - The folder is not fully self-contained for a full run, and SKILL.md and its references say so by pointing outside it: (a) the role-framing note in ../papercut-explainer/references/frame-breakdown.md (a sibling skill's file, 12,437 bytes at the pinned commit, not installed here); (b) storytelling.md sections 2, 4 and 6 point to ../../papercut-strategy/SKILL.md (audience level and the reel format), ../../papercut-hook/SKILL.md (the hook update) and ../../papercut-storyboard/SKILL.md (transitions and camera); (c) the prerequisite episode files described in the prerequisites above (the BRIEF in episodes/NN-slug/research/facts.md and the HOOK block in episodes/NN-slug/papercut/STORYBOARD.md, from papercut-strategy or papercut-hook) and my series' SERIES.md for audience level and region; and (d) the optional backdrop sheet library/papercut/public/bg/backdrops/_sheet.jpg via $PAPERCUT_LIB. If any of those is missing at use time, tell me and let me decide whether to install the sibling skills, provide the file, or do a reduced script-only run using what is in this folder (storytelling.md carries the full system, the phrase bank and two fully staged examples, and its section 7 is a copy-paste storytelling prompt for another tool). - Nothing is sent anywhere during installation itself; the only download at install time is the skill folder from GitHub at the pinned commit. No file is run during installation, no key is asked for, and there is no telemetry in the folder. - What leaves this machine at use time, disclosed here before it ever happens — using the skill is a separate step that only starts when I ask for a script and give you a source link: (a) research: you send the link I gave you (and any follow-on pages of that source) to your own web fetch or search tool and to the target site itself, and you use your browser tool to open the source pages and take screenshots of the key screens (2-4 of them) — so my source link, the search terms you derive from it, and the page requests go to your fetch/search/browser provider and to the source site, under your tools' own accounts and terms. Only facts with a source may be collected (name, maker, numbers, price, requirements, limitations, competitor numbers), and anything uncertain must be marked. If you have no such tools, do not improvise a substitute: ask me for the screenshots and facts instead, and nothing leaves that way. (b) Optional backdrop sheet: only if I have the full library locally and you resolve $PAPERCUT_LIB to choose sets, opening library/papercut/public/bg/backdrops/_sheet.jpg puts that local image into your context at about 0.7-1.5k tokens; if you are a cloud-hosted agent, that image goes to your model provider like any other context — open only what the step needs, as SKILL.md says. This install does not fetch the library. (c) The copy-paste prompt in references/storytelling.md section 7 is for me to paste into another tool myself if I choose: if I do, the source URL or notes I fill in go to that other tool, by my own action — never do that paste, or send the script or facts to any other tool, for me during installation or use. (d) Nothing else leaves: the writing itself (script, banners, staging table, humour pass) is done by you from the facts list, and its outputs are written locally to my episode folder: research/facts.md (the facts list, under the BRIEF), narration/script.json ([{"id": "l1_hook", "text": "..."}, ...], one entry per sentence or clause group), and the staging table in episodes/NN-slug/papercut/STORYBOARD.md under the HOOK block. No API key, account, or personal information is read or sent by the skill itself, and every number in the script, banners and rough UI must trace back to facts.md — never invent product behaviour, specs or numbers. - Do not run papercut-strategy, papercut-hook, papercut-storyboard or any other sibling skill during this install, do not fetch the backdrop sheet or the rest of the library, and do not start any research: those are my decisions, later, when I actually ask for a script. 3. When I say "yes, install", install it as a skill: copy the whole papercut-script folder into the agent's skills directory, in a folder named "papercut-script", keeping all 3 files (SKILL.md at its root and the two references in its references/ subfolder). A manual clone followed by copying the skills/papercut-script folder into ~/.claude/skills/ is the route described here; any other route is fine only as long as every installed file matches its fingerprint. 4. Verify with no network calls: frontmatter valid (name: papercut-script), all 3 files in place with matching fingerprints. There is nothing to run to verify. 5. Report what was installed, where, and what I still need to do myself (for example: provide a source link for the topic; make sure the earlier stages exist — the BRIEF in episodes/NN-slug/research/facts.md and the HOOK block in episodes/NN-slug/papercut/STORYBOARD.md, or install papercut-strategy and papercut-hook myself — plus my series SERIES.md; accept that research uses my agent's web fetch/search and browser tools, with the sends disclosed above, or provide screenshots and facts myself; if I want the backdrop sheet for choosing sets, have the full library locally and $PAPERCUT_LIB set; using the skill is a separate step that only starts when I ask for it, and the use-time sends disclosed above happen only then). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.