fix(integrations): drop invalid --model/--output-format flags from Vibe dispatch

[kartsan03]

Description

The Mistral Vibe CLI has no --model and no --output-format flag (checked on the latest release, 2.25.8). VibeIntegration inherits SkillsIntegration.build_exec_args(), which appends both, so a dispatched step exits 2 at argument parsing whenever either one ends up in the argv: a workflow step (or workflow default) that sets model:, or output_json=True.

Same shape as the Forge (#4667) and Amp (#4581) fixes:

• keep -p, Vibe's programmatic mode
• map output_json to Vibe's own --output json
• drop model: Vibe has no per-invocation model flag and picks its model from active_model in its config (or VIBE_ACTIVE_MODEL)

Vibe is one of the unverified integrations listed in #2416.

Reproduction on current main with vibe 2.25.8 on PATH (no API key configured) and this workflow in a project initialized with --integration vibe:

schema_version: "1.0"
workflow:
  id: "vibe-model-repro"
  name: "Vibe model repro"
  version: "1.0.0"
steps:
  - id: specify
    command: speckit.specify
    integration: vibe
    model: mistral-medium-latest
    input:
      args: "add a login page"

specify workflow run repro.yml:

vibe: error: unrecognized arguments: --model
Status: failed
Error: Command exited with code 2

With this change the step reaches Vibe itself and stops only at the missing credentials:

Error: Missing MISTRAL_API_KEY environment variable for mistral provider. ...
Status: failed
Error: Command exited with code 1

The JSON flag directly: vibe -p x --output-format json exits 2 with unrecognized arguments: --output-format, while vibe -p x --output json parses.

Testing

• Tested locally with uv run specify --help

• Ran existing tests with uv sync && uv run pytest

• Tested with a sample project (if applicable)

• New tests in tests/integrations/test_integration_vibe.py: prompt mode, --output json, model dropped, extra args. The last three fail on main and pass with the change.

• Full suite: 8536 passed, 214 skipped (Linux, Python 3.13)

• uvx ruff@0.15.0 check src tests: clean

• Sample project: the workflow above against main and this branch, vibe 2.25.8.

AI Disclosure

• I did not use AI assistance for this contribution
• I did use AI assistance (fill in the disclosure below)

AI disclosure: Claude Code (Claude Opus 5.5, max reasoning effort, autonomous agent mode) was used for drafting/refactoring the code change, the regression tests and this description.

[Copilot]

Copilot review overview

🟢 Approval recommended

The implementation matches Vibe 2.25.8’s CLI contract and includes focused positive and negative regression tests.

Review effort: Balanced
Findings: None

What changed in this PR

Corrects Vibe workflow dispatch to use supported CLI flags.

Changes:

• Maps JSON output to --output json.
• Drops unsupported per-invocation model flags.
• Adds regression coverage for prompt mode, JSON output, model omission, and extra arguments.

File	Description
src/​specify_cli/​integrations/​vibe/​__init__.py	Adds Vibe-specific execution arguments.
tests/​integrations/​test_integration_vibe.py	Covers corrected dispatch behavior.

---

💡 Configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[mnriem]

Thank you!