Skip to content
Back to search

building-with-typesafe-jev

MarketplaceWorks with Claude CodeGitHub

Unofficial skill for building with TypeSafe AI's Jev: typed decisions, calibrated confidence, and links to 150+ community projects, with a code sketch per pattern.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/aaddrick-building-with-typesafe-jev/badge.svg)](/listings/aaddrick-building-with-typesafe-jev)

HTML

<a href="/listings/aaddrick-building-with-typesafe-jev"><img src="/api/listings/aaddrick-building-with-typesafe-jev/badge.svg" alt="findsafeskills safety rating" /></a>

Install

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

These steps run a script from the network that findsafeskills does not scan.

<details> <summary><strong>Claude Code</strong></summary>

claude plugin marketplace add aaddrick/building-with-typesafe-jev
claude plugin install building-with-typesafe-jev@building-with-typesafe-jev

The skill loads on its own when you work on Jev code. To load it by hand, type:

/building-with-typesafe-jev:building-with-typesafe-jev

</details>

<details> <summary><strong>Claude Desktop, Cowork, and claude.ai</strong></summary>

**Step 1.** Open **Customize > Plugins**, select **Add**, then **Add marketplace**.

**Step 2.** Choose **Add from a repository**.

**Step 3.** Enter `aaddrick/building-with-typesafe-jev`. Leave **Sync automatically** on so the plugin updates when this repository does. Then select **Sync**.

**Step 4.** Select **Add** next to **Building with typesafe jev**.

**Step 5.** Claude confirms the plugin is installed. It also appears in the desktop app and Cowork on the same account.

</details>

<details> <summary><strong>Codex</strong></summary>

codex plugin marketplace add aaddrick/building-with-typesafe-jev
codex plugin add building-with-typesafe-jev@building-with-typesafe-jev

Start a new thread. Codex loads the skill when the task matches. To load it by hand, type:

$building-with-typesafe-jev:building-with-typesafe-jev

</details>

<details> <summary><strong>Antigravity CLI</strong></summary>

agy plugin install https://github.com/aaddrick/building-with-typesafe-jev

Check that it installed:

agy plugin list

Start a new session. Antigravity CLI loads the skill when the task matches. To load it by hand, type:

/building-with-typesafe-jev:building-with-typesafe-jev

Coming from Gemini CLI? If `agy plugin import gemini` brought this extension over, run the install command above anyway so the current copy replaces the imported one.

</details>

<details> <summary><strong>Muse (muse.ai)</strong></summary>

Muse loads skills from `~/workspace/skills/` on its own computer. Paste this command into a Muse chat and ask Muse to run it:

curl -fsSL https://raw.githubusercontent.com/aaddrick/building-with-typesafe-jev/main/scripts/install_muse.sh | bash

The script copies the skill folder there and rewrites the `SKILL.md` header into the shape Muse reads. Start a new chat. Muse loads the skill when the task matches. To update, run the command again.

</details>

<details> <summary><strong>Muse Code</strong></summary>

Clone the repository:

git clone https://github.com/aaddrick/building-with-typesafe-jev.git

Install the skill for every project:

muse skills install building-with-typesafe-jev/skills/building-with-typesafe-jev --scope user

Check that it installed:

muse skills list

Start a new session. Muse Code loads the skill when the task matches. To load it by hand, type:

/building-with-typesafe-jev

</details>

<details> <summary><strong>Any other agent that reads SKILL.md</strong></summary>

Copy the `skills/building-with-typesafe-jev/` folder into your agent's skills folder. Keep the whole folder. `SKILL.md` links to the files beside it.

</details>

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

.claude-plugin/marketplace.json · Manifest · 24 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. {
  2. "$schema": "https://www.schemastore.org/claude-code-marketplace.json",
  3. "name": "building-with-typesafe-jev",
  4. "description": "Unofficial skill for building with TypeSafe AI's Jev: typed decisions, calibrated confidence, and links to 150+ community projects, with a code sketch per pattern.",
  5. "owner": {
  6. "name": "aaddrick",
  7. "url": "https://github.com/aaddrick"
  8. },
  9. "plugins": [
  10. {
  11. "name": "building-with-typesafe-jev",
  12. "description": "Teaches Claude the Jev API, when to use Choice, Score, or Noul, how to route on confidence, and where to find prior art for a design before building it.",
  13. "source": "./",
  14. "category": "development",
  15. "keywords": [
  16. "typesafe",
  17. "jev",
  18. "system-one",
  19. "classification",
  20. "structured-output"
  21. ]
  22. }
  23. ]
  24. }
README.md · README · 403 lines · 1 finding marked

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <p align="center">
  2. <strong>Building with TypeSafe Jev</strong><br>
  3. <em>Typed decisions with calibrated confidence, for your coding agent.</em><br>
  4. <em>Links to 150+ community projects, sorted by shape, with a code sketch for each.</em>
  5. </p>
  6. <p align="center">
  7. <a href="LICENSE"><img src="https://img.shields.io/github/license/aaddrick/building-with-typesafe-jev?style=flat" alt="License"></a>
  8. <a href=".github/workflows/checks.yml"><img src="https://img.shields.io/github/actions/workflow/status/aaddrick/building-with-typesafe-jev/checks.yml?label=checks&style=flat" alt="Checks"></a>
  9. </p>
  10. <p align="center">
  11. <a href="https://www.linkedin.com/in/aaddrick/">Connect on LinkedIn!</a>
  12. </p>
  13. <p align="center">
  14. <strong>English</strong> ·
  15. <a href=".github/readme/README.zh-CN.md">简体中文</a> ·
  16. <a href=".github/readme/README.ja.md">日本語</a> ·
  17. <a href=".github/readme/README.ko.md">한국어</a> ·
  18. <a href=".github/readme/README.vi.md">Tiếng Việt</a> ·
  19. <a href=".github/readme/README.pt-BR.md">Português (BR)</a> ·
  20. <a href=".github/readme/README.it.md">Italiano</a>
  21. </p>
  22. > [!NOTE]
  23. > This is an unofficial, community skill. It is not made, reviewed, or endorsed by TypeSafe AI. TypeSafe publishes its own skill at [typesafe-ai/skills](https://github.com/typesafe-ai/skills). See [How this differs from the official skill](#how-this-differs-from-the-official-skill).
  24. Coding agents treat Jev like one more chat model. This skill teaches them to design for it: typed questions, calibrated confidence, and links to 150+ community projects, sorted by how they work, with a code sketch for each pattern. It installs in Claude Code, Codex, Antigravity CLI, Muse, and Muse Code.
  25. [Jev](https://docs.typesafe.ai/introduction) is a [System One](https://docs.typesafe.ai/concepts/system-one) model. It does not write text. You send it content and a set of typed questions, and it answers each one with a value and a calibrated probability, usually in 100 to 200 ms:
  26. - **Choice** picks one option from a list. Example: route a ticket to billing, shipping, or support.
  27. - **Score** places the content on a scale you describe. Example: grade a pull request from "ignores the spec" to "meets the spec".
  28. - **Noul** gives the probability that a yes/no statement is true. Example: "this shell command deletes files outside the project."
  29. ## Install
  30. <details>
  31. <summary><strong>Claude Code</strong></summary>
  32. ```bash
  33. claude plugin marketplace add aaddrick/building-with-typesafe-jev
  34. ```
  35. ```bash
  36. claude plugin install building-with-typesafe-jev@building-with-typesafe-jev
  37. ```
  38. The skill loads on its own when you work on Jev code. To load it by hand, type:
  39. ```
  40. /building-with-typesafe-jev:building-with-typesafe-jev
  41. ```
  42. </details>
  43. <details>
  44. <summary><strong>Claude Desktop, Cowork, and claude.ai</strong></summary>
  45. **Step 1.** Open **Customize > Plugins**, select **Add**, then **Add marketplace**.
  46. <img src=".github/assets/plugin-marketplace/step-1.png" alt="The Plugins page in Customize with the Add menu open. An amber box and arrow point at Add marketplace." width="100%">
  47. **Step 2.** Choose **Add from a repository**.
  48. <img src=".github/assets/plugin-marketplace/step-2.png" alt="The Add marketplace dialog. An amber box and arrow point at Add from a repository." width="100%">
  49. **Step 3.** Enter `aaddrick/building-with-typesafe-jev`. Leave **Sync automatically** on so the plugin updates when this repository does. Then select **Sync**.
  50. <img src=".github/assets/plugin-marketplace/step-3.png" alt="The Add marketplace dialog with aaddrick/building-with-typesafe-jev in the URL field and Sync automatically on. An amber box and arrow point at the URL field, the toggle, and the Sync button." width="100%">
  51. **Step 4.** Select **Add** next to **Building with typesafe jev**.
  52. <img src=".github/assets/plugin-marketplace/step-4.png" alt="The Discover list showing Building with typesafe jev. An amber box and arrow point at its Add button." width="100%">
  53. **Step 5.** Claude confirms the plugin is installed. It also appears in the desktop app and Cowork on the same account.
  54. <img src=".github/assets/plugin-marketplace/step-5.png" alt="The Building with typesafe jev plugin page with a notice that it is installed and ready to use. An amber box and arrow point at the notice." width="100%">
  55. </details>
  56. <details>
  57. <summary><strong>Codex</strong></summary>
  58. ```bash
  59. codex plugin marketplace add aaddrick/building-with-typesafe-jev
  60. ```
  61. ```bash
  62. codex plugin add building-with-typesafe-jev@building-with-typesafe-jev
  63. ```
  64. Start a new thread. Codex loads the skill when the task matches. To load it by hand, type:
  65. ```
  66. $building-with-typesafe-jev:building-with-typesafe-jev
  67. ```
  68. </details>
  69. <details>
  70. <summary><strong>Antigravity CLI</strong></summary>
  71. ```bash
  72. agy plugin install https://github.com/aaddrick/building-with-typesafe-jev
  73. ```
  74. Check that it installed:
  75. ```bash
  76. agy plugin list
  77. ```
  78. Start a new session. Antigravity CLI loads the skill when the task matches. To load it by hand, type:
  79. ```
  80. /building-with-typesafe-jev:building-with-typesafe-jev
  81. ```
  82. Coming from Gemini CLI? If `agy plugin import gemini` brought this extension over, run the install command above anyway so the current copy replaces the imported one.
  83. </details>
  84. <details>
  85. <summary><strong>Muse (muse.ai)</strong></summary>
  86. Muse loads skills from `~/workspace/skills/` on its own computer. Paste this command into a Muse chat and ask Muse to run it:
  87. ```bash
  88. curl -fsSL https://raw.githubusercontent.com/aaddrick/building-with-typesafe-jev/main/scripts/install_muse.sh | bash
    • Note: Pipes a downloaded script directly into a shell (curl | sh / wget | bash)
    • Line 131 on GitHub
  89. ```
  90. The script copies the skill folder there and rewrites the `SKILL.md` header into the shape Muse reads. Start a new chat. Muse loads the skill when the task matches. To update, run the command again.
  91. </details>
  92. <details>
  93. <summary><strong>Muse Code</strong></summary>
  94. Clone the repository:
  95. ```bash
  96. git clone https://github.com/aaddrick/building-with-typesafe-jev.git
  97. ```
  98. Install the skill for every project:
  99. ```bash
  100. muse skills install building-with-typesafe-jev/skills/building-with-typesafe-jev --scope user
  101. ```
  102. Check that it installed:
  103. ```bash
  104. muse skills list
  105. ```
  106. Start a new session. Muse Code loads the skill when the task matches. To load it by hand, type:
  107. ```
  108. /building-with-typesafe-jev
  109. ```
  110. </details>
  111. <details>
  112. <summary><strong>Any other agent that reads SKILL.md</strong></summary>
  113. Copy the `skills/building-with-typesafe-jev/` folder into your agent's skills folder. Keep the whole folder. `SKILL.md` links to the files beside it.
  114. </details>
  115. ## Set up an API key (optional, recommended)
  116. The skill works without a key. With `TYPESAFE_API_KEY` set in the agent's shell, the agent can check its design against the live API before the code reaches your project. That catches wrong field names, and questions Jev reads differently than you meant. Each call costs a fraction of a cent.
  117. <details>
  118. <summary><strong>Create, store, and troubleshoot a key</strong></summary>
  119. <br>
  120. <details>
  121. <summary><strong>Create a key</strong> (four steps in the TypeSafe console)</summary>
  122. **Step 1.** Sign in at [console.typesafe.ai](https://console.typesafe.ai/) and open **API Keys** in the sidebar.
  123. <img src=".github/assets/api-key/step-1.png" alt="The TypeSafe console home page. An amber box and arrow point at API Keys in the left sidebar." width="100%">
  124. **Step 2.** Click **Create key** at the top right.
  125. <img src=".github/assets/api-key/step-2.png" alt="The API keys page. Existing keys are blurred. An amber box and arrow point at the Create key button at the top right." width="100%">
  126. **Step 3.** Name the key after where it will live, such as the machine or the agent. Then click **Create key**.
  127. <img src=".github/assets/api-key/step-3.png" alt="The Create API key dialog with the name my-coding-agent typed in. An amber box and arrow point at the name field and the Create key button." width="100%">
  128. **Step 4.** Copy the key now. The console shows it once. If you lose it, create a new one and revoke the old one.
  129. <img src=".github/assets/api-key/step-4.png" alt="The API key created dialog. The key value is masked. An amber box and arrow point at the Copy button." width="100%">
  130. </details>
  131. <details>
  132. <summary><strong>Store the key</strong> (macOS, Linux, Windows)</summary>
  133. Keep the key in its own file, readable only by you, and export it as `TYPESAFE_API_KEY`. Agents often start shells without a terminal, so each section puts the key where those shells can see it. Pick your system.
  134. <details>
  135. <summary><strong>macOS</strong> (zsh, the default shell)</summary>
  136. Save the key to a private file:
  137. ```bash
  138. mkdir -p ~/.config/typesafe && umask 077 && printf 'export TYPESAFE_API_KEY=%s\n' 'YOUR_KEY' > ~/.config/typesafe/env
  139. ```
  140. Load it from `~/.zshenv`. Every zsh reads that file, including shells that agents start without a terminal. `~/.zshrc` is read only by interactive shells.
  141. ```bash
  142. echo '[ -f ~/.config/typesafe/env ] && . ~/.config/typesafe/env' >> ~/.zshenv
  143. ```
  144. Open a new terminal and check it. The command prints the length of the key, not the key:
  145. ```bash
  146. echo ${#TYPESAFE_API_KEY}
  147. ```
  148. </details>
  149. <details>
  150. <summary><strong>Linux with bash</strong> (Ubuntu, Debian, Mint, Fedora, Arch)</summary>
  151. Save the key to a private file:
  152. ```bash
  153. mkdir -p ~/.config/typesafe && umask 077 && printf 'export TYPESAFE_API_KEY=%s\n' 'YOUR_KEY' > ~/.config/typesafe/env
  154. ```
  155. Load it from the **top** of `~/.bashrc`. Ubuntu, Debian, Mint, and Arch start `~/.bashrc` with a line that stops early when no terminal is attached. A line below that guard never runs for agent shells. The top of the file is safe on every distribution:
  156. ```bash
  157. sed -i '1i [ -f ~/.config/typesafe/env ] \&\& . ~/.config/typesafe/env' ~/.bashrc
  158. ```
  159. Open a new terminal and check it:
  160. ```bash
  161. echo ${#TYPESAFE_API_KEY}
  162. ```
  163. </details>
  164. <details>
  165. <summary><strong>Linux with zsh</strong></summary>
  166. Follow the macOS steps. zsh reads `~/.zshenv` the same way on Linux.
  167. </details>
  168. <details>
  169. <summary><strong>Linux with fish</strong></summary>
  170. fish reads every file in `~/.config/fish/conf.d/`, with or without a terminal:
  171. ```fish
  172. mkdir -p ~/.config/fish/conf.d; and echo 'set -gx TYPESAFE_API_KEY YOUR_KEY' > ~/.config/fish/conf.d/typesafe.fish; and chmod 600 ~/.config/fish/conf.d/typesafe.fish
  173. ```
  174. ```fish
  175. string length -- $TYPESAFE_API_KEY
  176. ```
  177. </details>
  178. <details>
  179. <summary><strong>Windows</strong> (PowerShell)</summary>
  180. Store the key as a user environment variable. New terminals and apps see it. Terminals that are already open do not:
  181. ```powershell
  182. [Environment]::SetEnvironmentVariable('TYPESAFE_API_KEY', 'YOUR_KEY', 'User')
  183. ```
  184. Open a new terminal and check it:
  185. ```powershell
  186. $env:TYPESAFE_API_KEY.Length
  187. ```
  188. **WSL:** Windows variables do not reach WSL by default. Inside WSL, follow the Linux with bash steps.
  189. </details>
  190. <details>
  191. <summary><strong>Desktop apps and IDE extensions</strong></summary>
  192. An app you start from the dock, the start menu, or a desktop launcher does not read your shell files.
  193. - **Windows:** the user environment variable above already covers these apps.
  194. - **Linux (systemd):** add the line `TYPESAFE_API_KEY=YOUR_KEY` to `~/.config/environment.d/typesafe.conf`, then log out and back in.
  195. - **macOS:** start the app from a terminal, or set the variable in the app's own settings. macOS has no simple per-user file that desktop apps read.
  196. </details>
  197. </details>
  198. <details>
  199. <summary><strong>If the agent cannot see the key</strong></summary>
  200. Ask the agent to run `echo ${#TYPESAFE_API_KEY}` (or `$env:TYPESAFE_API_KEY.Length` on Windows). If it prints `0` or nothing, check these in order:
  201. - **You started the agent before you stored the key.** Agent shells copy the environment of the program that started them. Quit the agent and start it from a new terminal.
  202. - **You started the agent from the dock, the start menu, or an IDE.** Those apps do not read your shell files. See "Desktop apps and IDE extensions" above.
  203. - **Codex filters the environment.** If `~/.codex/config.toml` sets `include_only` under `[shell_environment_policy]`, add `TYPESAFE_API_KEY` to it. If it sets `ignore_default_excludes = false`, Codex drops every variable with `KEY` in its name. Remove that line.
  204. - **WSL.** Windows variables do not reach WSL. Store the key inside WSL with the Linux steps.
  205. </details>
  206. </details>
  207. ## How it was tested
  208. We gave a coding agent six Jev tasks, such as a support-ticket triage function and an approval gate for shell commands. Each task ran 10 times with this skill, with no skill, and with the official skill, each arm in its own isolated container. The agent had the docs but no API key, so graders checked the code it wrote. Checks that need judgment went to three LLM judges from three providers (Claude Opus, GPT-6 Sol, Kimi K3), and the majority decided.
  209. | | No skill | This skill | Official skill |
  210. |---|---:|---:|---:|
  211. | Average score | 0.65 ± 0.05 | **0.96 ± 0.02** | 0.77 ± 0.04 |
  212. Score is the share of checks passed, averaged over 10 runs per task.
  213. **Where the skill made the difference** (runs out of 10 that passed):
  214. | The agent's code... | No skill | This skill | Official skill |
  215. |---|---:|---:|---:|
  216. | counted in code, instead of asking Jev for a number | 1 | 8 | 0 |
  217. | named a field of the input in its questions | 0 | 10 | 1 |
  218. | pinned or logged the Jev model version | 0 | 10 | 0 |
  219. | gave the department list a catch-all option | 3 | 10 | 6 |
  220. | kept more than one path while walking 1,200 categories | 1 | 10 | 7 |
  221. | kept a plain-code backstop for destructive commands | 7 | 10 | 2 |
  222. | read `score` as a position from 0 to n-1 | 9 | 10 | 5 |
  223. [See every check →](evals/docs/results.md#every-check)
  224. Each of these rows beats no skill, the official skill, or both by more than chance at 95%.
  225. **More detail:**
  226. - [evals/README.md](evals/README.md): results by case, and how to run the suite
  227. - [evals/docs/results.md](evals/docs/results.md): every check, each judge's scores, cost, tokens, and method
  228. - [evals/docs/cases.md](evals/docs/cases.md): the six tasks and what each check looks for
  229. - [evals/docs/harness.md](evals/docs/harness.md): how a run works, the isolated containers, and how to keep a batch
  230. - [evals/docs/lessons.md](evals/docs/lessons.md): what broke while the suite was being built
  231. ## What is inside
  232. The skill loads in layers, so the agent reads only what the task needs.
  233. | File | What it holds | When the agent reads it |
  234. |---|---|---|
  235. | `SKILL.md` | Which primitive to pick, 11 design rules, how to use probabilities and confidence, common mistakes | Every Jev task |
  236. | `api-reference.md` | HTTP API, Python and JavaScript SDKs, limits, errors, environment variables | When it writes the code |
  237. | `patterns.md` | The 4 official patterns and techniques from 18 cookbooks, with their thresholds | When it designs a workflow |
  238. | `prior-art/INDEX.md` | A map from "what I want to build" to a shape, plus ideas that failed | Before it designs something new |
  239. | `prior-art/*.md` | 11 shape files: a code sketch, field lessons, and linked projects | One or two per design |
  240. ## The prior-art library
  241. Most catalogs sort projects by industry. This library sorts them by implementation shape. A game bot, a drone, and a trading bot share one shape: a control loop. Sorted that way, the three share one code sketch and one set of field lessons. The 11 shapes:
  242. - **[Control loops](skills/building-with-typesafe-jev/prior-art/control-loops.md)**: games, drones, robots, markets.
  243. - **[Select from candidates](skills/building-with-typesafe-jev/prior-art/select-from-candidates.md)**: browser and phone agents, tool calling without an LLM, extraction, routers.
  244. - **[Gates](skills/building-with-typesafe-jev/prior-art/gates.md)**: tool-call approval, "done" checks, CI, money, content.
  245. - **[Stream filters](skills/building-with-typesafe-jev/prior-art/stream-filters.md)**: slop filters, moderation, email, logs, bulk labels.
  246. - **[Ranking and matching](skills/building-with-typesafe-jev/prior-art/ranking-and-matching.md)**: rerankers, entity matching, graph and taxonomy walks.
  247. - **[Judges and evals](skills/building-with-typesafe-jev/prior-art/judges-and-evals.md)**: rubric judges, trace grading, code review as triage.
  248. - **[Incremental and real-time](skills/building-with-typesafe-jev/prior-art/incremental-realtime.md)**: dubbing, voice, keystroke-driven interfaces.
  249. - **[Agent context and memory](skills/building-with-typesafe-jev/prior-art/agent-context-memory.md)**: compaction, memory gates, memory expiry, effort control.
  250. - **[Pairing with an LLM](skills/building-with-typesafe-jev/prior-art/llm-pairing.md)**: planner and actor, verify-then-escalate, distillation.
  251. - **[Answers as data](skills/building-with-typesafe-jev/prior-art/research-and-features.md)**: features for classical models, research instruments, benchmarks.
  252. - **[Embedding in infrastructure](skills/building-with-typesafe-jev/prior-art/embedding-in-infrastructure.md)**: SQL functions, vector databases, CI hooks, Home Assistant.
  253. The index also lists the ideas that failed in the field: chess, code review as the only reviewer, perception, and calibration taken on trust. A failed attempt saves the next builder from repeating it.
  254. ## How this differs from the official skill
  255. The [official TypeSafe skill](https://github.com/typesafe-ai/skills) is one file of design guidance. For API details, it sends the agent to the live docs on every task. The two skills have different names and do not conflict, so you can install both, though the evals did not test them together.
  256. This skill holds more inside the skill itself: exact API shapes, cookbook thresholds, community failure modes, and the prior-art library. The agent can design without a network round trip and see what others built first.
  257. Jev changes fast. The skill is a snapshot of [docs.typesafe.ai](https://docs.typesafe.ai/llms.txt) and the community on 2026-09-25, and it tells the agent the live docs win on any conflict. If a fact is wrong, please open an issue with a link to the source.
  258. ## Credits
  259. The facts about the API come from TypeSafe AI's public documentation and cookbooks. The prior art comes from the builders who published their projects, from the Hacker News community, and from these indexes: [awesome-jev-typesafe](https://github.com/valentynkit/awesome-jev-typesafe), [JevDirectory](https://www.jevdirectory.org/resources), [awesome-jev-use-cases](https://github.com/walidboulanouar/awesome-jev-use-cases), and [awesome-typesafe-jev](https://github.com/AbdelStark/awesome-typesafe-jev). Each shape file links to the original projects.
  260. TypeSafe, Jev, and System One are names of TypeSafe AI. This project uses them only to say what the skill is for.
  261. ## License
  262. MIT. See [LICENSE](./LICENSE).
scripts/install_muse.sh · Bundled script · 80 lines · 1 finding marked
  1. #!/usr/bin/env bash
  2. # Install the skill into Muse (muse.ai), Meta's personal agent.
  3. #
  4. # Muse loads skills from ~/workspace/skills/<folder>/SKILL.md on its own
  5. # computer. It reads a narrower frontmatter than Claude Code or Codex: the
  6. # name must use underscores, and both values must be one-line quoted strings.
  7. # This script copies the skill folder there and rewrites the frontmatter into
  8. # that shape. The rest of the folder is copied unchanged.
  9. #
  10. # Run it inside Muse's computer (ask Muse to run the command in the README):
  11. #
  12. # curl -fsSL https://raw.githubusercontent.com/aaddrick/building-with-typesafe-jev/main/scripts/install_muse.sh | bash
    • Note: Pipes a downloaded script directly into a shell (curl | sh / wget | bash)
    • Line 12 on GitHub
  13. #
  14. # Environment:
  15. # MUSE_SKILLS_DIR where Muse reads skills (default: ~/workspace/skills)
  16. # MUSE_SKILL_SOURCE a local skills/building-with-typesafe-jev folder to
  17. # install instead of downloading the latest release
  18. #
  19. # Run it again to update. It replaces the installed copy.
  20. set -euo pipefail
  21. SKILL=building-with-typesafe-jev
  22. REPO=aaddrick/building-with-typesafe-jev
  23. DEST_ROOT="${MUSE_SKILLS_DIR:-$HOME/workspace/skills}"
  24. DEST="$DEST_ROOT/$SKILL"
  25. work="$(mktemp -d)"
  26. trap 'rm -rf "$work"' EXIT
  27. if [ -n "${MUSE_SKILL_SOURCE:-}" ]; then
  28. src="$MUSE_SKILL_SOURCE"
  29. else
  30. curl -fsSL "https://codeload.github.com/$REPO/tar.gz/refs/heads/main" | tar -xz -C "$work"
  31. src="$work/${REPO#*/}-main/skills/$SKILL"
  32. fi
  33. if [ ! -f "$src/SKILL.md" ]; then
  34. echo "error: no SKILL.md in $src" >&2
  35. exit 1
  36. fi
  37. stage="$work/stage/$SKILL"
  38. mkdir -p "$(dirname "$stage")"
  39. cp -R "$src" "$stage"
  40. rm -rf "$stage/__pycache__"
  41. # Rewrite the frontmatter. Keep the description, drop every other key, and
  42. # quote both values. The body below the closing --- is copied as is.
  43. skill_md="$stage/SKILL.md"
  44. if [ "$(head -n 1 "$skill_md")" != "---" ]; then
  45. echo "error: $skill_md has no frontmatter" >&2
  46. exit 1
  47. fi
  48. end="$(awk 'NR > 1 && $0 == "---" { print NR; exit }' "$skill_md")"
  49. if [ -z "$end" ]; then
  50. echo "error: $skill_md frontmatter never closes" >&2
  51. exit 1
  52. fi
  53. desc="$(sed -n "2,$((end - 1))p" "$skill_md" | sed -n 's/^description:[[:space:]]*//p')"
  54. if [ -z "$desc" ]; then
  55. echo "error: $skill_md has no one-line description" >&2
  56. exit 1
  57. fi
  58. desc="${desc//\\/\\\\}"
  59. desc="${desc//\"/\\\"}"
  60. name="${SKILL//-/_}"
  61. {
  62. printf -- '---\nname: "%s"\ndescription: "%s"\n---\n' "$name" "$desc"
  63. tail -n "+$((end + 1))" "$skill_md"
  64. } > "$work/SKILL.md"
  65. mv "$work/SKILL.md" "$skill_md"
  66. mkdir -p "$DEST_ROOT"
  67. rm -rf "$DEST"
  68. mv "$stage" "$DEST"
  69. echo "Installed $name into $DEST"
  70. echo "Start a new Muse chat so it picks up the skill."

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.