jev skillA
jev is agent-read markdown (skill) from rashedint32/jev-mcp: Get typed, calibrated judgments from the connected `jev` MCP tools when the same semantic decision repeats across many items — triaging files, ranking candidates, checking claims against their sources, scoring a diff against each requirement — or when you need a calibrated number to threshold on. Prefer an ordinary answer for a single judgment in conversation. Not for judgments that ship inside the user's application; those belong in SDK code, not in a tool call..
Indexed from public GitHub and served as immutable, content-addressed versions. Install it pinned to an exact SHA-256 with the mdr CLI, and every file is verified against the hash recorded here before it reaches your agent. The deterministic audit below grades the latest version, and the same file always earns the same grade.
What the file says
# Jev judgments over MCP Jev is a System One model. It returns a typed answer and a calibrated probability distribution, never prose. The `jev` MCP server exposes it as six tools. This skill covers **when to call those tools and how to shape the call**. For the underlying theory — primitive semantics, state design, composition patterns — use the `typesafe-ai` skill or the live docs at <https://docs.typesafe.ai>. Do not duplicate that material here. If the `jev_*` tools are not in your tool list, this skill does not apply. Call `jev_models` to confirm the key works before assuming a failure is your own. ## Three-way test Decide which of these you are in before calling anything. **Answer it yourself** when the judgment happens once, in conversation, and you can explain your reasoning. That is most judgments. An ordinary answer carries an argument the user can push back on. Jev returns a number and a label, which is weaker in dialogue. Do not reach for a tool to look rigorous. **Call the tool** when the same judgment repeats across many items, or when you need a calibrated number to threshold on. Triaging forty files. Ranking twenty candidates. …
Read the whole file at its exact version.
How to install
mdr add rashedint32/jev-mcp/jev@git:20260920.ad00bf8mdr add rashedint32/jev-mcp/jev@sha256:d579ac9b3bee5e8cPin to a label to follow the author's releases, or to a sha256 to freeze the exact bytes forever. Either way the resolved hash is written to mdr.lock, and mdr install reproduces it on any machine.
[](https://markdownregistry.com/a/art_y4wxs44yam4ctwv2)
1 badge views in 30 days
Versions
| version | committed | commit | size | audit | |
|---|---|---|---|---|---|
| git:20260920.ad00bf8 latest | 2026-09-20 | ad00bf8 | 10,164 B | A | view · diff |
| git:20260917.27603c1 | 2026-09-17 | 27603c1 | 8,480 B | A | view · diff |
| git:20260917.9893777 | 2026-09-17 | 9893777 | 7,960 B | A | view · diff |
| git:20260917.1690489 | 2026-09-17 | 1690489 | 5,234 B | A | view |
Audit of the latest version
- pass: Frontmatter block present
- pass: Frontmatter declares a name
- pass: Frontmatter declares a description
- pass: Size between 200 bytes and 200 KB (10164 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
rashedint32/jev-mcp · 7 stars · license MIT · pushed 2026-09-23 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_y4wxs44yam4ctwv2 GET https://markdownregistry.com/api/v1/resolve?ref=rashedint32/jev-mcp/jev GET https://markdownregistry.com/api/v1/blob/d579ac9b3bee5e8c343969e791d2ab09a3828e20edd206bde37d5b0b48436dfc
Your agent does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.