paleo
ðĶī paleo â personal token-saving skill collection.
Open source Open in the app JSON README (API)
About
ðĶī paleo â personal token-saving skill collection.
Details
- Kind
- Plugins
- Topic
- No topic detected
- Publisher
- mocasus
- Origin
- gemini
- Category
- ferramentas
- Version
- 2.5.0
- Stars
- 20
- Last push
- 2026-07-25T18:11:03Z
- Repository state
- ativo
- Language
- Python
- License
- MIT
- Added
- 2026-08-30 14:13:39
- Updated
- 2026-08-30 14:13:39
- Origin id
mocasus/paleo
README
<div align="center">  [](https://github.com/mocasus/paleo/releases) · [](https://github.com/mocasus/paleo/blob/main/BENCHMARK.md) · [](LICENSE) · [](https://www.python.org) · [](https://github.com/mocasus/paleo) · [](BENCHMARK.md) · [](https://agentskills.io) # ðĶī paleo > Token-saving skills for LLM agents â cut output & context tokens without choking the model. [Why](#why-paleo) · [See it](#see-it-before--after) · [Features](#features) · [Skills](#skills) · [Combos](#recommended-combos) · [Quick Start](#quick-start) · [Benchmarks](#benchmarks) · [Comparison](#comparison) · [Tips](#tips--triggers) · [Install](#installation) </div> ## Why paleo? - **Tokens cost money and latency.** Every trimmed token means faster, cheaper inference. - **One-size-fits-all prompting fails.** Sometimes you want terse output, sometimes a hard budget, sometimes just no fluff. paleo gives each as a separate, well-scoped skill. - **Skills stay minimal.** Every `SKILL.md` is written terse on purpose â loading one costs less context than a long prompt. ## See it (Before / After) paleo compresses *delivery*, not meaning. Code, commands, and technical terms stay byte-exact. | Normal agent | ðĶī paleo | |---|---| | "The re-render happens because you create a new object literal on every render. That inline object is a fresh reference each time, so React sees a changed prop and re-renders. Wrap it in `useMemo` to keep a stable reference." | "New object each render â new ref â re-render. Wrap in `useMemo`. Stable ref = no re-render." | | "To authenticate requests, add middleware that checks the token on each request and returns 401 if it is missing or expired." | "Add auth middleware. Check token per request. 401 if missing/expired." | > [!NOTE] > paleo keeps technical accuracy at 100% â it drops filler, not facts. ## Features - [x] **Modular & composable** â load one skill or all seven, mix per task. - [x] **Output + context savings** â ~50â70% fewer output tokens (median ~54% on a 6-task sample â see [BENCHMARK.md](./BENCHMARK.md)), plus proactive context trimming. - [x] **Auto-detect** â `paleo-auto` watches session state and enables the right skills automatically. No thinking required. - [x] **Production-safe** â compresses output and context only; never rewrites your code. - [x] **Hard token budget** â `paleo-budget` caps spend and summarizes the tail. - [x] **Cross-agent** â open Agent Skills standard: Claude Code, Codex, Gemini CLI, Qwen Code, OpenCode, Cursor, GitHub Copilot, Cline, Windsurf + **40+ agents** via `npx skills add`. - [x] **Zero-setup triggers** â plain English phrases, no slash commands to register. - [x] **Low overhead** â each `SKILL.md` is intentionally terse, so loading stays cheap. - [x] **Open & extensible** â drop in your own token-saving skills. ## Skills | Skill | What it does | Trigger example | |---|---|---| | `paleo` | Terse output mode â cut output tokens ~50â70%, keep code/terms exact. | `paleo mode` · `be brief` · `save tokens` | | `paleo-auto` | ð Zero-touch auto-detection â watches session & enables the right skills automatically. | `paleo-auto` · `auto paleo` · `enable auto-save` | | `paleo-budget` | Hard token budget per task â cap spend, summarize if exceeded. | `budget 2000` · `stay under 2000 tokens` | | `paleo-trim-context` | Proactively trim/summarize context to save tokens without losing task state. | `trim context` (auto on long sessions) | | `paleo-converse` | Condense old chat turns + merge duplicate messages; keep last N verbatim. | `condense chat` · `compress conversation` · `paleo-converse N=8` | | `paleo-summary` | Tight intisari of bulky tool output / logs / diffs / dumps. | `tldr` · `condense this` · `summarize output` | | `paleo-json` | Minify & compact structured/JSON output, stay parseable. | `compact json` · `minify` | ## Recommended combos `paleo` is the base â keep it on. Layer the rest by what you're doing: | Situation | Combo | Why | |---|---|---| | Daily driver (long / chatty sessions) | `paleo` + `paleo-trim-context` | Base + automatic context hygiene. | | History piling up | + `paleo-converse` | Condense + merge duplicate turns once a session gets long. | | Debugging / bulky tool output | `paleo` + `paleo-summary` + `paleo-json` | Logs â intisari; JSON â minified. | | Tight cost / hard limit | `paleo` + `paleo-budget` (+ `paleo-trim-context`) | Hard ceiling + shrink context first. | | Max savings (all on) | all seven skills | Overkill daily, but safe for extreme thrift. | > [!TIP] > You rarely need every skill at once. `paleo` + `paleo-auto` is the default for most users â auto-detect handles the rest. For manual control: `paleo` + `paleo-trim-context` is the baseline; add `paleo-converse` for messy chats, `paleo-summary` / `paleo-json` for heavy tool output, and `paleo-budget` only when a hard cap is required. ## Quick Start ```bash # 1. Clone the collection git clone https://github.com/mocasus/paleo.git # 2. Claude Code â one plugin bundles all 7 skills claude plugin marketplace add https://github.com/mocasus/paleo claude plugin install paleo@paleo # 3. Any agent via the open Agent Skills registry (installs to 40+ clients) npx skills add mocasus/paleo ``` Then just talk to your agent â no command to register: > `paleo mode` · `save tokens` · `budget 2000` · `trim context` ## Benchmarks Real, reproducible numbers â not hand-waved claims. | Model | Tasks | Median output savings | Mean | |---|---|---|---| | `claude-sonnet-4.5` | 6 | **53.8%** | 45.1% | Full method, per-task table, and the runnable harness are in [BENCHMARK.md](./BENCHMARK.md). Rerun on your own stack: ```bash export IDROUTER_API_KEY=your_key python3 bench/benchmark.py --model claude-sonnet-4.5 ``` > [!TIP] > Savings are task-dependent: biggest on verbose generative work (code, walkthroughs, comparisons â 54â79%), smaller on already-compact factual answers. paleo also cuts *context* tokens via `paleo-trim-context`, a layer a terse-persona prompt cannot reach. ## Comparison paleo is often compared with two other token-saving approachesa **terse-persona system prompt** and **Ponytail** (a code-reuse coding skill). Here is how they differ. | Dimension | ðĶī paleo | Terse-persona prompt | Ponytail | |---|---|---|---| | **Form** | 7 composable skills | Single system prompt (persona) | Single coding skill / workflow | | **What it targets** | Output tokens **+** context **+** conversation turns | Output tokens only | Volume of code the agent writes (+ MCP caching) | | **Granularity** | Per-task, mix & match | One mode | One workflow | | **Touches your code** | â No (output/context only) | â No | â ïļ Yes â refactors / reuses code | | **Context & reasoning savings** | â `paleo-trim-context` | â None | â Partial (caching) | | **Hard budget** | â `paleo-budget` | â | â | | **Cross-agent** | â 40+ agents (open standard) | â Portable prompt, but monolithic | â Claude Code skill | | **Activation** | Plain phrases | Edit system prompt | Install + invoke skill | | **Reasoning-model safe** | â Never compresses thought | â Can *raise* tokens (e.g. +3% on Opus) | â | | **Known risk** | None (output-only) | Can fight "expand" heuristics; may *raise* tokens on reasoning models | Refactor can change behavior | | **Open benchmark** | â Reproducible harness | â Claim only | â Claim only | > [!TIP] > **They're complementary, not rivals.** Ponytail cuts the *code you have to write*; paleo cuts the *tokens in the conversation*. Terse-persona prompts proved a terse prompt helps output â paleo takes that same idea and makes it modular, adds context-trimming and a hard budget, and drops the persona gimmick. Use Ponytail for code-heavy work and paleo for chatty, long sessions. ## FAQ <details> <summary><b>Prompt dipangkas, context dipotong, kualitas gak bakal sama dong?</b></summary> Gak lebih jelek â malah sering lebih bagus. paleo bukan potong buta, dia buang *redundansi* (ulang-ulang, filler, boilerplate hasil tool, whitespace), bukan info esensial. Constraint, error, code, keputusan tetap utuh. Prompt lo gak diapa-apain â yang di-trim itu context kerja agent (hasil tool berulang, riwayat convo). Context bersih = model fokus ke signal, bukan lost track karena noise. Benchmark: median **53.8%** token turun, kualitas task gak drop. Trade-off jujur: budget ekstrem (token cap rendah banget) bisa turun kualitas, tapi setting *optional*. Intinya: lebih murah & cepet, kualitas tetap. </details> <details> <summary><b>System prompt itu buat bikin model bagus, klo dipangkas banyak yg ilang dong?</b></summary> paleo gak pernah sentuh system prompt. System prompt = aturan main, utuh 100%. Instruksi yang define behavior gak diapa-apain â paleo jalan *setelah* system prompt ke-load, cuma kerja di context dinamis. Yang di-trim itu context kerja (tool output berulang, convo kepanjangan), bukan instruksi. System prompt kecil dibanding noise yang numpuk dari tool output. Plus bisa whitelist bagian yang mau dijaga. Efeknya kebalik: context bersih bikin model *lebih* patuh ke system prompt. Aturan main tetep nempel. </details> <details> <summary><b>Bisa ngerusak code / formatting gak?</b></summary> Gak. paleo melindungi code block, structured output (JSON/table), dan error message secara default. Yang di-compress cuma prose bertele-tele & tool output redundant. Kalo masih ragu, bisa whitelist file/section tertentu biar 100% gak ke-trim. </details> <details> <summary><b>Ini butuh API key atau service eksternal?</b></summary> Enggak. paleo murni teknik prompt/context â gak ada server, gak ada API call, gak ada dependensi luar. Skill-nya tinggal di-load ke agent lo, jalan di lokal. No overengineering. </details> <details> <summary><b>Token savings-nya beneran kelihatan di billing?</b></summary> Kelihatan, terutama di session panjang & agent loop. Tiap token yang gak dikirim = gak dibayar. Benchmark kita median **53.8%** turun di context + output. Di agent yang muter 20+ tool call, itu selisih gede per run. </details> <details> <summary><b>Works di semua model/provider?</b></summary> Iya. paleo model-agnostic â kerja di level prompt & context, bukan di model tertentu. Claude, GPT, Gemini, GLM, Qwen, lokal â semua bisa. Sifatnya instruksi, bukan fine-tune. </details> <details> <summary><b>Beda sama auto-compaction bawaan agent (Claude compaction, dll)?</b></summary> Compaction bawaan itu generic & reaktif (baru jalan pas context mau penuh, sering blind truncation). paleo proaktif + selektif: jaga info esensial, buang redundansi, configurable, ada safety net (whitelist). Plus paleo juga ngurus output verbosity & tool-result summarization, bukan cuma convo history. </details> <details> <summary><b>Cara enable/disable per task?</b></summary> Trigger pakai natural language (`skip preamble`, `ringkas output`, `trim context`) â gak perlu slash command. Mau matiin? Tinggal gak dipanggil, atau cabut skill dari agent. No global lock-in. </details> ## Tips & Triggers > paleo activates from natural-language triggers â no slash command to register. Type the trigger, the skill loads and applies. <details> <summary>Activation & switches (plain phrases)</summary> **paleo** â terse output - On: `paleo mode` · `be brief` · `terse` · `compress output` · `save tokens` - Level: `paleo full` (default) · `paleo lite` · `paleo ultra` - Off: `stop paleo` · `normal mode` **paleo-budget** â token cap - On: `budget 2000` · `stay under 2000 tokens` · `token limit` - Off: `no budget` · `unlimited` **paleo-trim-context** â auto on long sessions; `trim context` to force. **paleo-auto** â `paleo-auto` · `auto paleo` â off: `disable paleo-auto` · `manual paleo` **paleo-converse** â `condense chat` · `compress conversation` · `paleo-converse N=8` **paleo-summary** â `tldr` · `condense this` · `summarize output` **paleo-json** â `compact json` · `minify` **Combo:** `paleo` + `paleo-budget` = max savings. Add `paleo-trim-context` on long sessions, `paleo-converse` on chatty ones, `paleo-summary` for bulky tool output. </details> ## Installation <details><summary><b>ðĩ Claude Code</b></summary> ```bash claude plugin marketplace add https://github.com/mocasus/paleo claude plugin install paleo@paleo ``` </details> <details><summary><b>ðĒ Codex</b></summary> ```bash npx skills add mocasus/paleo ``` </details> <details><summary><b>ðĄ Gemini CLI</b></summary> ```bash mkdir -p ~/.gemini/skills && cp -r skills/* ~/.gemini/skills/ ``` </details> <details><summary><b>ðĢ Hermes Agent</b></summary> ```bash hermes skills install mocasus/paleo # Or copy skills manually cp -r skills/paleo* ~/.hermes/skills/ ``` </details> <details><summary><b>âŦ Cursor</b></summary> ```bash npx skills add mocasus/paleo ``` </details> <details><summary><b>ð· GitHub Copilot</b></summary> ```bash npx skills add mocasus/paleo ``` </details> <details><summary><b>ð Windsurf</b></summary> ```bash npx skills add mocasus/paleo ``` </details> <details><summary><b>ðĶ Universal (any agent)</b></summary> ```bash git clone https://github.com/mocasus/paleo.git # copy skills/paleo*/ into your agent's skills directory ``` </details> All 7 skills load automatically â `paleo`, `paleo-trim-context`, `paleo-auto`, `paleo-budget`, `paleo-converse`, `paleo-summary`, `paleo-json`. > Full per-agent steps in [INSTALL.md](./INSTALL.md). See real compression numbers in [BENCHMARK.md](./BENCHMARK.md). ## Hermes Integration > paleo is battle-tested on [Hermes Agent](https://github.com/NousResearch/hermes-agent) by [@mocasus](https://github.com/mocasus) â first user & case study. ```bash # Install via Hermes skills manager hermes skills install mocasus/paleo # Or by path cp -r skills/paleo* ~/.hermes/skills/ ``` Then in your Hermes chat (Telegram, WhatsApp, etc.): ``` > paleo ðĶī paleo full â terse output, code-first > build a REST API with FastAPI [terse, code-first response â no preamble, no filler] > paleo-auto ðĶī paleo-auto: watching session... enabled paleo + trim-context (23 turns) > budget 2000 ðĶī paleo-budget: 2000 output tokens, hard mode ``` **Hermes tips:** - Start with `paleo` â instant token savings on every reply. - `paleo-auto` for sessions >15 turns (watches context fill & enables the right skills). - Combine `paleo` + `budget` for expensive models via provider routing. - `paleo-converse` kicks in when your agent loop hits 60%+ context cap. ## Custom Skills paleo is open â wire your own token-saving skills: 1. `skills/<your-name>/SKILL.md` with `name` + `description` frontmatter. 2. Add the skills directory to `.claude-plugin/plugin.json` â the `skills` field is a path **string** (e.g. `"./skills/"`), not an array. Gemini + other agents pick skills up natively; no extra manifest needed. 3. Bump version badge (this file + footer) + plugin `version`. 4. Commit + push. No repo edit needed â just drop any `SKILL.md` into your agent's skills dir (e.g. `~/.hermes/skills/<name>/`). paleo loads whatever it finds under `skills/`. ## User Stats > Share your numbers, get listed. PR your monthly token savings to this table. | User / Team | Agent | Tokens/month saved | Skills | |---|---|---|---| | *[Add yours â](https://github.com/mocasus/paleo/issues/new?title=stats)* | â | â | â | **How to measure:** 1. Use agent 1 week without paleo â note token usage from provider dashboard 2. Enable `paleo` (or `paleo-auto`) 1 week â note new usage 3. Diff à 4 = estimated monthly savings ## Contributing Contributions are welcome â new token-saving skills, better triggers, or benchmark data. - Open an issue describing the skill or improvement. - Keep `SKILL.md` files terse (they load into context). - Add `name` + `description` frontmatter and register in both plugin manifests. - Bump the version badge and `version` fields before opening a PR. ## License MIT â see [LICENSE](./LICENSE). --- <div align="center"> ## Sponsors <a href="https://kliqo.co"><img src="./assets/kliqo-banner.jpg" alt="Kliqo.co" width="420"></a> **Kliqo.co** sponsors paleo · <a href="https://kliqo.co">kliqo.co</a> ðĶī paleo · v2.5.0 · MIT </div>