agentleFS
Sign inStart free
amirkiarafieiamirkiarafiei/subagent-cli-skills/skills/mistral-vibe/SKILL.md

mistral-vibe

Skill14 starsChanged 10 days agoMITView on GitHub

Use Mistral Vibe CLI as a subagent. Lets the main agent prompt Mistral Vibe CLI from the terminal in headless mode with the `vibe` command, and send the goal, the decisions and the scope with the task. Use when the user asks to delegate work to Mistral Vibe CLI, when a task needs a second independent agent, or when a plan needs a fresh perspective.

  • Pipes a download into a shell

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Secrets or policy-sensitive work: Do not pipe credentials into a subagent.

    line 48
  2. Distinctive:

    Low return on investment: If writing the handoff costs more than doing the edit, do the edit.

    line 50
  3. Distinctive:

    Treat every call as stateless delegation: the subagent keeps nothing between calls, so each command must carry its own state.

    line 61
  4. Distinctive:

    Write the prompt as a transfer of shared state, not as a title.

    line 71
  5. Distinctive:

    Prefer sequential delegations with explicit carry-over.

    line 82
  6. Distinctive:

    After the subagent returns:

    line 96
  7. Distinctive:

    Every vendor has its own terminal rules.

    line 104
  8. Distinctive:

    Exit code 0 is not success: If stdout is empty, treat the run as failed and read stderr.

    line 108
  9. Very specific:

    Keep all seven sub-headings, in this order, even if the answer is "none".

    line 119
  10. Distinctive:

    If the user names a model, use it.

    line 179

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Mistral Vibe CLI (subagent / task delegation)

<!-- ============================================================================= GLOBAL HALF — sections 1 to 5. These sections are identical in every skill in this repository. Copy them without any change. Do not put a vendor name or a flag in them. A change here must be made in all skills at the same time. ============================================================================= -->

What is it

This skill teaches your agent to use an agent from another vendor as a subagent.

Almost every coding agent has a headless mode. Headless mode runs the agent from the terminal with one command and returns the answer to stdout. This skill gives the command pattern for one such agent, and the protocol to transfer enough context with the task.

You stay the main agent. You keep the plan, the decisions and the conversation with the user. The subagent does one bounded task and reports back.

When to delegate

  • The user asks for it. The user names another agent, or asks for a second opinion.
  • Large or multi-step work. The task covers several files, phases or checkpoints.
  • Code review with a fresh mind. Another agent reviews the work with a different lens and no memory of the decisions that produced it.
  • Planning that needs different ideas. A second agent proposes options that you did not consider.
  • An internal council. Several subagents answer the same question, and you compare the answers.
  • Deep research. The other CLI has web search, web extraction or browser automation.
  • Background investigation. A long task runs in the background while you continue in the foreground.

When not to delegate

  • Small or single-step tasks. One or two edits, or a short explanation, are faster in this thread.
  • Tight feedback loops. The user wants quick back-and-forth in one conversation.
  • Secrets or policy-sensitive work. Do not pipe credentials into a subagent. Redact first.
  • Context you already hold. Repeating the whole plan in a prompt adds no value.
  • Low return on investment. If writing the handoff costs more than doing the edit, do the edit.
  • Work that needs high precision and full judgement. A subagent without your context will break a task that depends on a detail only this session knows.

Delegation and context transfer protocol

An isolated subagent saves tokens, but it splits the story. The subagent does not see this conversation. It has no memory of the goal you agreed with the user, the decisions you already made, or the paths you agreed not to touch. A poor handoff causes misread tasks, wrong assumptions about the stack or the style, and wasted edits.

Treat every call as stateless delegation: the subagent keeps nothing between calls, so each command must carry its own state. A good handoff happens when you keep these practices in mind:

  • Shared state handoff. The prompt carries the original goal, the decisions already made, and the explicit scope, to bridge the subagent's missing session history.
  • Contextual preservation. The high-volume work stays in the isolated process, so the bulk never enters this conversation and no "split story" forms from missing history.
  • Result reconciliation. What comes back is folded into this thread afterwards, so there is a single source of truth.

Write the prompt as a transfer of shared state, not as a title. Include all six fields:

IncludeWhy
GoalThe same objective as the user, not only the immediate micro-task.
DecisionsFramework, patterns, naming, "use X not Y" — anything that would otherwise be guessed wrong.
ScopeThe absolute paths and modules to touch, and the areas to leave alone. A relative path may not resolve where you expect.
ConstraintsPerformance, accessibility, compatibility, review gates, "no new dependencies".
VerificationThe exact command the subagent must run and pass before it returns.
OutputFor example: "report only, no edits", "apply edits with a minimal diff", "list the files changed".

Prefer sequential delegations with explicit carry-over. Parallel runs diverge unless every run gets the same briefing.

Run the subagent in the mode that does the work without asking for approval. Headless mode has nobody to answer a permission prompt. Any mode that stops to ask will stall, or return an empty answer with a success exit code. Each CLI names this mode differently — take the flag from the vendor card below.

Never call a subagent in plan mode. Many CLIs have a plan or read-only mode. That mode makes the other agent write a plan for itself, which is not what you asked for: you want its work or its answer, and you keep the planning. Plan mode also waits for someone to approve that plan, so the run comes back with nothing. If you want no file changes, keep the auto-approving mode and write "report only, no edits" in the Output field. Do not use plan mode to make a run read-only.

After the subagent returns:

  • Report to the user. Give a short summary. Do not paste long logs unless the user asks.
  • Reconcile the context. Record the decisions it made, the files it changed, and the open risks, so this session stays the single source of truth.

CLI failure modes

Every vendor has its own terminal rules. A headless run fails quietly more often than it fails loudly.

  • Wrap the call in an external timeout. A headless CLI can stall before its own timeout starts.
  • Exit code 0 is not success. If stdout is empty, treat the run as failed and read stderr. A permission the CLI could not ask for, and a prompt that never arrived, both look like success.
  • Never carry a flag habit from one CLI to another. The same short flag means different things in different tools. In OpenCode, -p is --password: passing a prompt to it empties the message and the run waits on stdin forever. Confirm every flag against the CLI's own --help.
  • An unrecognized-flag error means this skill is stale, not that the task is impossible. Read --help, continue with the flags that exist, and tell the user which line in this file is wrong.
<!-- ============================================================================= VENDOR HALF — section 6. Everything below is specific to this CLI. Keep all seven sub-headings, in this order, even if the answer is "none". Take every flag from the installed binary (`<cli> --help`), not from memory and not from another skill in this repository. ============================================================================= -->

Vendor card

Binary and prompt form

  • Binary: vibe
  • Install: curl -LsSf https://mistral.ai/vibe/install.sh | bash, or uv tool install mistral-vibe (both from the official repo README, https://github.com/mistralai/mistral-vibe).
  • Prompt form: flag. The prompt is the value of --prompt "prompt", e.g. vibe --prompt "Analyze the codebase" --max-turns 5 --output json. Passing --prompt puts Vibe in programmatic mode, which skips the interactive chat UI entirely and disables interactive tools (e.g. ask_user_question).

Headless and output flags

NeedFlag
Non-interactive prompt--prompt "prompt"
Output format--output text|json|streaming (text default)
Turn cap--max-turns <N>
Cost cap--max-price <dollars> — documented as "indicative only", a soft/approximate limit, not a hard guarantee
Token cap--max-tokens <N>
Deadline--time-limit <seconds> — stops the run; the session can be resumed
Restrict tool set--enabled-tools <list> — exact names, globs (bash*), or regex via re: prefix (e.g. re:^serena_.*$)
Resume--continue/-c (most recent session) or --resume <SESSION_ID> (specific session)

Approvals and permissions

Approval is controlled by --agent <name>. The documented built-in agents are ask (asks before tool calls), accept-edits (the default: auto-approves file edits, needs approval for the rest), smart-approve, and auto-approve (approves every tool call — the vendor says to use it only in a sandboxed environment). A fifth agent, lean, is for Lean 4 work only. Vibe has no built-in plan agent. The flags --auto-approve and --yolo also exist and approve every tool call. An admin can set enable_auto_approve = false, and then both flags are ignored.

A headless run does not auto-approve by default. A vibe --prompt … run uses the default agent (accept-edits). Nobody can answer an approval prompt, so Vibe denies every tool call that needs approval — shell commands included. For a run that must act, pass --agent auto-approve explicitly. Finer-grained control exists via per-tool "always"/"ask" settings, bash allow/deny lists, and --trust (grants temporary folder trust for the current invocation only, non-persistent).

Exit codes (vendor docs, not reproduced here): 0 the agent finished, 1 the run could not start, 2 model API or runtime failure, 3 stopped early (limit reached, stop, or the model declined). Use --time-limit <seconds> for a deadline. An older community report described a run that stalled on an internal plan-confirmation step. It is unverified for the current release, so keep an external timeout around the run anyway.

Models and how to list them

No vibe models subcommand or --list-models flag is documented. Selection is --model <alias> (it starts a new session on that model and does not change active_model); model IDs can be a -latest alias (e.g. a "medium" or "small" tier) or a dated/version-pinned snapshot made by replacing -latest with a date/version suffix — prefer a pinned snapshot for reproducible delegations. An interactive /model slash command exists to select the active model, but whether it enumerates the full catalog or only opens a picker is not confirmed.

If the user names a model, use it. Otherwise check the vendor's own documentation and Artificial Analysis for current names and pricing before picking one.

Command pattern

vibe --prompt "GOAL: [goal] | DECISIONS: [decisions] | SCOPE: [paths] | CONSTRAINTS: [constraints] | VERIFICATION: [test_command] | OUTPUT: [format]" \
  --agent auto-approve --model <identifier> --output text --time-limit 2700 2>&1

Prompt examples

  • Implement: vibe --prompt "GOAL: [goal] | DECISIONS: [decisions] | SCOPE: [paths] | VERIFICATION: [test_command] | OUTPUT: [format]" --agent auto-approve --model <identifier>
  • Investigate (report only): vibe --prompt "GOAL: Map how [feature] works | SCOPE: [paths] | CONSTRAINTS: do not edit any file | OUTPUT: concise file:line map" --agent accept-edits --model <small-tier identifier> (shell calls are denied under accept-edits; use --agent auto-approve if the investigation needs to run commands)
  • Bounded run: vibe --prompt "GOAL: [task] | VERIFICATION: [test_command] | OUTPUT: [format]" --agent auto-approve --max-turns 15 --max-price 2

Docs and reference

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free