Harnesses

Mistral Vibe

Mistral's CLI. Everything under ~/.vibe and trust per folder. MCP servers as TOML tables and a read-only mode down to read_file.
IDvibe15 / 1714 paths · env + project
Harness

Mistral Vibe

  • vibe

Mistral's CLI. MCP servers as TOML tables, trust per folder, read-only down to read_file.

Invoke
4 of 6 · official
Models
no listing
MCP
toml · vibe
Agents file
~/.vibe/AGENTS.md
Inferred 0 / 14

Access

GetgetHarness("vibe")
CLIharnesses paths vibe --json
Tryvibe in the explorer

Mistral Vibe is Mistral's coding CLI, and the binary is vibe. It installs as a Python tool, but the launcher just hands over to a bundled Rust binary. Neat.

Everything lives under ~/.vibe/, and VIBE_HOME moves all of it. User config is ~/.vibe/config.toml, a project adds .vibe/config.toml. A VIBE_<KEY> variable overrides one key. Hooks sit in hooks.toml, in both places.

Is your project folder trusted? If not, Vibe skips its config, its skills and its AGENTS.md. A headless run warns, then answers without them. Trusted folders go into ~/.vibe/trusted_folders.toml, and --trust covers a single run. The registry won't add that flag for you.

agents sync links ~/.vibe/AGENTS.md for your own instructions. Project ones are every AGENTS.md from the working directory up to the trusted root, the closest winning.

Skills come from .vibe/skills/ and .agents/skills/ in a project, ~/.vibe/skills/ and ~/.agents/skills/ at home. Skills sync owns ~/.vibe/skills/, so after a full sync Vibe finds each skill twice and quietly keeps the first. Custom slash commands are skills too, so there's no prompt template folder.

Sessions get a directory each under ~/.vibe/logs/session/unified/<session-id>/. Inside: meta.json, journal JSONL files and content chunks.

MCP servers are [[mcp_servers]] tables in config.toml, each one naming itself. vibe mcp add writes the user file. Transports are stdio, http and streamable-http, no sse. A name with a space or a dot gets rewritten on load, so the library refuses it before writing. Better than a server that answers to a name you never gave it.

vibe --prompt=<text> is the headless mode. The prompt stays glued to --prompt=, because -p reads a prompt that starts with a dash as a flag.

--output json prints the session messages as one JSON array at the end. --model takes a catalog alias like mistral-medium-latest, and a wrong one fails with the list of real ones. --thinking carries effort. An unknown word gets a warning, then no thinking level at all.

Runs use default_agent, which is accept-edits unless changed. Nobody answers approval prompts headless, so anything that would ask gets denied.

Read-only passes --disabled-tools with a pattern that hits everything but read_file. Disabled tools add up across config layers and agent profiles, so no profile can switch one back on.

--enabled-tools wasn't enough: a custom profile with its own list replaced it and wrote files. run_typescript, search_tool_functions and wait stay. That sandbox has no files, no fetch and no process. Bash, edits, subagents, skills and MCP are all denied.

There's no advisor mode. Switch every tool off and the model still sees the whole toolbox. It then spends its turns knocking on locked doors.

Every command Vibe starts gets VIBE_SESSION=1. That one variable is all detectHarness needs. A .vibe folder is the project marker.

Capabilities and invocation

  • mcp: yes
  • vision: yes
  • audio: no
  • video: no
  • tools: yes
  • streaming: yes

Binary vibe, evidence official. The prompt rides on --prompt= because -p reads one that starts with a dash as a flag. Runs use default_agent, accept-edits unless changed, and a headless run denies every call that would ask. Project AGENTS.md and .vibe/ load only in a trusted folder. Read-only runs disable every tool but read_file: bash, edits, subagents, skills, connectors and MCP tools are denied, and run_typescript stays as a sandbox without files or network. --enabled-tools does not qualify, since an agent profile's own enabled_tools replaces it. Checked on 3.1.3 against a config and an agent profile that allow everything. No advisor: with every tool disabled the model is still offered them and burns its turns on denials. Model selection appends --model {model}.

Callinvoke(prompt, options)4 of 6 modes
  • advisor{}rejected
  • advisor · json{ structured: true }rejected
  • read-only{ readOnly: true }--prompt={prompt} --disabled-tools re:^(?!read_file$).*
  • read-only · json{ readOnly: true, structured: true }--prompt={prompt} --disabled-tools re:^(?!read_file$).* --output json
  • agent{ tools: true }--prompt={prompt}
  • agent · json{ tools: true, structured: true }--prompt={prompt} --output json

Paths

Templates as the registry stores them. ~, ${HOME}, ${TMPDIR} and %VAR% expand in resolve(); entries tagged with a platform are dropped on the others. Open in the explorer to see them expanded for a home directory of your choice.

PathsgetHarness("vibe")14 templates · 5 groups

config

  • ~/.vibe/config.tomluserofficialall

    Base directory overridable with VIBE_HOME. VIBE_<KEY> variables override single keys.

  • .vibe/config.tomlprojectofficialall

    The closest one walking up from the cwd, read only in a trusted folder.

  • ~/.vibe/trusted_folders.tomluserofficialall

    Folders whose project config, skills and AGENTS.md Vibe loads.

  • ~/.vibe/.envuserofficialall

    API keys and provider credentials.

  • ~/.vibe/agents/userofficialall

    Custom agent profiles as TOML, picked with --agent. .vibe/agents/ in a project.

sessions

  • ~/.vibe/logs/session/unified/<session-id>/dataofficialall

    meta.json, journal/*.jsonl and chunks/ per session. session_logging.save_dir moves it.

instructions

  • AGENTS.mdprojectofficialall

    Every one from the cwd up to the trusted root, the closest winning.

  • ~/.vibe/AGENTS.mduserofficialall

    Global user-level instructions.

skills

  • .vibe/skills/projectofficialall
  • .agents/skills/projectofficialall
  • ~/.vibe/skills/userofficialall
  • ~/.agents/skills/userofficialall

    Read by default, so the Codex skills target covers Vibe too.

hooks

  • .vibe/hooks.tomlprojectofficialall
  • ~/.vibe/hooks.tomluserofficialall

Empty

commandspromptTemplatestempThat is not the same as the feature being missing.

Env overrides

Variables the harness reads to move a root. The path is the default while the variable is unset; resolve() applies a set one to temp.

EnvenvOverrides1 variables
  • VIBE_HOME~/.vibeconfig · sessions · instructions · skills · hooks

MCP servers

FilemcpConfigs2 files
  • ~/.vibe/config.tomluser · toml · vibe · mcp_servers
  • .vibe/config.tomlproject · toml · vibe · mcp_servers

Detection

CalldetectHarness()env, then project

envVars

  • VIBE_SESSION

projectMarkers

  • .vibe

Persistence

  • TOML - Config, hooks, trusted folders, agent profiles. (official)
  • JSONL - Session journals. (official)
  • JSON - Session metadata and content chunks. (official)

syncAgentsFiles links ~/.vibe/AGENTS.md to the master bundle.