Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add nataliacorrea03/claude-code-skills --skill swipegit clone --depth 1 https://github.com/nataliacorrea03/claude-code-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/nataliacorrea03/claude-code-skills/swipe)<a href="https://agentmods.dev/skills/nataliacorrea03/claude-code-skills/swipe"><img src="https://agentmods.dev/badge/skills/nataliacorrea03/claude-code-skills/swipe/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/nataliacorrea03/claude-code-skills/swipe"><img src="https://agentmods.dev/badge/skills/nataliacorrea03/claude-code-skills/swipe.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00188 | $0.02659 |
| Opus 5.5 | $0.00075 | $0.01064 |
| Sonnet 5.5 | $0.00038 | $0.00532 |
| Haiku 4.5 | $0.00019 | $0.00266 |
Grade C, and why
swipe scanned grade C with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2mo ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
9. **Report** one plain line, e.g. "41 trashed, 12 kept. 5 replies queued (3 yours, 2 I drafted), say 'send it' to send them." Then **shut down the server** (kill the background process) and **wipe the run dir**: `rm -rf How it starts
The opening of the file, as written. The whole thing — 131 lines — stays where its author put it; the contents beside it link to each section on GitHub.
swipe
A swipe-to-triage deck for digital cleanup. You point it at a source, it serves a card deck to your phone over your local WiFi, you swipe through the whole pile in one sitting, and the actions run as one reversible batch when you are done.
Why this exists
A folder with 300 screenshots is a wall. Cleaning it the normal way means 300 separate decisions, each with its own friction (open it, look at it, decide, find the next one), and that friction is exactly what task paralysis feeds on. So the folder just sits there, growing.
This collapses every decision into one flick of a thumb. One card, one swipe, next card. The deck shows up on your phone, so you can do it on the couch instead of staring at a file manager. Each swipe is its own tiny hit of done, and the momentum carries you through the pile that you would otherwise never start. Nothing is destructive in the moment: files go to the Trash (recoverable), text replies are staged for you to send later. The payoff is fast and visible, and you can undo.
Design rules that make it work, do not break them:
- Push, not pull. It loads the whole pile and hands it to you. It never makes you come back to fetch more.
- No guilt, no streaks, no counters that shame you. Just the deck and a running position.
- Reversible. Trash, not delete. Staged replies, not sent messages.
- One uninterrupted session. See below.
- Never make someone swipe blind. Every card should tell you enough to actually decide. See Previews below.
The engine is source-agnostic
The server and deck do not know or care what they are showing. They render whatever a manifest describes. A "source" is just an adapter script that emits that manifest. v1 ships two adapters (Files and iMessage); adding a new source (Gmail, Photos, a download queue, anything) is one new adapter that emits the same schema. That is what "point it at anything" means here.
Manifest schema (the contract)
{
"source": "files",
"title": "Downloads cleanup",
"left": { "label": "Trash", "kind": "trash" },
"right": { "label": "Keep", "kind": "keep" },
"reply": false,
"items": [
{
"id": "0",
"type": "image | file | text",
"label": "headline (file/image card)",
"sub": "subline, e.g. size and date, or a timestamp",
"path": "/abs/path (files: needed for thumbnail + action)",
"thumb": true,
"preview": "inline text preview shown on the card (files or text)",
"sender": "headline for a text card",
"handle": "who to reply to (text)",
"row": 123
}
]
}
What ships with it
12 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
- .gitignore 25 B
- adapters/files.py 3.7 KB runs code
- adapters/imessage.py 1.6 KB runs code
- deck.html 8.4 KB
- server.py 6.2 KB runs code
- tests/test_files_adapter.py 3.5 KB runs code
- tests/test_imessage_adapter.py 2.5 KB runs code
- tests/test_server.py 3.0 KB runs code
- tests/test_trash_files.py 2.5 KB runs code
- trash_files.py 1.3 KB runs code
- vendor/LICENSE-qrcodejs.txt 1.2 KB
- vendor/qrcode.js 20 KB runs code
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2mo ago First seen · 131 lines · 188 tokens per session scan C 5d2676c997ea
swipe is a skill published in the GitHub repository nataliacorrea03/claude-code-skills (4 stars, last pushed 3mo ago), licensed MIT. It adds 188 tokens to every session and 2,659 once invoked, about $0.0008 per session on Opus 5.5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
swipe compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| swipenataliacorrea03/claude-code-skills | Skill | 4 | 3mo ago | MIT | C | 188 | Claude Code, Codex |
| apple-remindersNousResearch/hermes-agent | Skill | 262k | today | MIT | A | 17 | Claude Code, Codex |
| apple-notesNousResearch/hermes-agent | Skill | 262k | today | MIT | A | 17 | Claude Code, Codex |
| boxNousResearch/hermes-agent | Skill | 262k | today | MIT | A | 14 | Claude Code, Codex |
| obsidianNousResearch/hermes-agent | Skill | 262k | today | MIT | A | 19 | Claude Code, Codex |
| channels-setupCopilotKit/CopilotKit | Skill | 39k | today | MIT | A | 141 | Claude Code |
| mantis-configuregoogle/mantis | Skill | 2.7k | 3d ago | Apache-2.0 | A | 77 | Claude Code, Codex |