v1.3.0

Full Changelog: v1.2.0...v1.3.0

v1.2.0: Video analysis mode

What's new

Video analysis mode

Analyze video files and YouTube URLs using Gemini's multimodal capabilities — scene breakdowns, ad quality reviews, and content verification.

# Local file
/midflight --video ./ad-v3.mp4

# YouTube URL
/midflight --video https://youtube.com/watch?v=abc123

# With a specific question
/midflight --video ./ad-v3.mp4 Does this match the storyboard?

Default output is a structured Markdown table with 6 columns: Timestamp, Visual Narrative, Text & Graphics, Cinematography & Editing, Audio & Soundscape, and Emotional Beat — plus a symbolism/color grading summary and ad evaluation.

Video mode auto-switches to Gemini regardless of your configured provider. Local files are sandboxed via a temp staging directory (max 20MB, Gemini CLI limit).

Bug fixes

• Fix silent failure when config keys are missing in ~/.config/mid-flight/config (#3)
• Remove temp file cleanup that triggers permission prompts (#4)

Full changelog: v1.1.0...v1.2.0

v1.1.0

mid-flight v1.1.0

On-demand consultation with Codex or Gemini — without leaving your Claude Code session.

Highlights

• Consult / Implement routing — Claude automatically classifies your query intent. Consult mode gets advice; implement mode enables file edits and verification. Defaults to consult when uncertain.
• Dual provider support — Route queries to Codex or Gemini CLI via ~/.config/mid-flight/config
• Preflight dependency checks — Verifies CLI tools are installed before querying, with actionable install URLs on failure
• macOS-compatible temp files — Uses TMPDIR-aware mktemp patterns that work across platforms
• Debug logging — Duration and query metadata logged to stderr with [mid-flight] prefix

Install

claude plugin marketplace add Abeansits/mid-flight
claude plugin install mid-flight@mid-flight

Usage

/midflight should we use WebSockets or SSE here?
/midflight

Companion to pre-flight.