Install
$ agentstack add skill-igapyon-miku-media-proc-skills-miku-media-proc ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
miku-media-proc
This skill helps Codex perform media processing work in this repository.
Use it when the user explicitly names miku-media-proc, asks to apply this skill, or asks for media processing workflows that belong to this repository.
Do not use it for unrelated general coding work, generic writing, or ordinary file inspection unless media processing is the main task.
Environment Assumptions
Assume the working environment has these tools installed and available:
ffmpegand related FFmpeg command-line tools.- Node.js.
- Python.
- DaVinci Resolve Studio.
- DaVinci Resolve scripting/API access, usable from automation code.
Do not spend effort installing these tools unless the user explicitly asks for setup or troubleshooting.
Operational Defaults
- Use
ffprobeas the default metadata-inspection command for local video and
audio sources before making project, timeline, frame-rate, or export decisions.
- Preserve source quality during working steps. Prefer stream copy, remuxing,
original media, or lossless/high-quality intermediates over repeated lossy re-encoding. Use WAV/PCM for audio intermediates when reliability matters and the file size is acceptable.
- When timeline creation, editing, or final export is needed and no other editor
is specified, prefer DaVinci Resolve Studio.
- For Resolve API automation, assume the user starts Resolve Studio manually,
then connect through the API. Use Python as the default direct Resolve API scripting language.
- For new Resolve projects, inspect source media first, set timeline resolution
and timelineFrameRate before timeline creation, then import, assemble, and verify placement. Decide horizontal vs vertical delivery before project or timeline creation. If timelinePlaybackFrameRate remains 24, report it instead of repeatedly forcing the setting.
- For Resolve timeline assembly, consider preparing separate video-only and
audio-only assets when source streams are unstable, audio needs WAV normalization, narration is added separately, or troubleshooting benefits from independent tracks.
- When creating task-specific automation scripts, record the script path,
purpose, executed actions, and generated or modified Resolve project/timeline names in a task memo under repository-root temp/.
- For 60fps-family game or motion-dense footage, preserve the source fps family
for output unless there is a concrete reason to convert.
- For local Zundamon narration, use the user-started VOICEVOX Engine HTTP API
at http://127.0.0.1:50021; if the user writes VOICEBOX, treat it as the same local VOICEVOX Engine workflow unless they clarify otherwise. Verify /version and /speakers before synthesis. Use Zundamon normal voice as the default narration voice unless the user requests a different speaker or style. Use the local HTTP API when possible, and use Python for API-driven narration scripts by default.
- For generated audio or video playback checks, prefer VLC media player and use
browser playback only as a supplemental check.
- For YouTube publishing, prepare and verify the upload candidate file locally,
but recommend that the user uploads it manually in YouTube Studio. Automated YouTube upload is not recommended for this skill at this time.
Core Workflow
- Clarify the media task: input files, desired output format, quality target,
constraints, and whether destructive changes are allowed.
- Ask what kind of video or media the sources represent when it affects the
workflow, such as game footage, screen recording, live action, commentary, tutorial, or social-media content.
- Inspect source media metadata before changing files when local media files
are available. Use FFmpeg-family tools for media inspection, and use ffprobe as the default metadata-inspection tool.
- When timeline creation, editing, or delivery export is needed and the user
has not specified another editor, guide the workflow toward DaVinci Resolve Studio first.
- After source files are available and media information has been inspected,
prompt the user to proceed to assembly editing in DaVinci Resolve Studio when a timeline deliverable is intended.
- For final video deliverables, confirm whether the program assembly should
include a title card, end card, or other front/back matter.
- Prefer reproducible commands and scripts over one-off manual steps.
- Write generated files under a task-specific output directory unless the user
asks for an exact destination.
- Verify output existence, duration or dimensions where relevant, and basic
metadata before reporting completion.
- When a media-processing problem is diagnosed and solved while using this
skill, record the problem, background, decision, and resolution in the project root DECISIONS.md; see [references/integration/decisions-log.md](references/integration/decisions-log.md).
References
- [index.json](index.json): generated inventory for this skill package.
- [references/integration/workflow.md](references/integration/workflow.md): general media-processing workflow, task routing, tool defaults, and work memo rules.
- [references/integration/decisions-log.md](references/integration/decisions-log.md): rules for recording solved media-processing decisions in project
DECISIONS.md. - [references/integration/frame-rate-policy.md](references/integration/frame-rate-policy.md): source frame-rate preservation and conversion policy.
- [references/davinci/resolve-24fps-recovery-case.md](references/davinci/resolve-24fps-recovery-case.md): DaVinci Resolve 24fps failure-and-recovery case for unstable smartphone or game-recorded media.
- [references/youtube/youtube-upload-fps.md](references/youtube/youtube-upload-fps.md): frame-rate guidance for YouTube upload targets.
- [references/youtube/youtube-format-duration-guidance.md](references/youtube/youtube-format-duration-guidance.md): YouTube normal video vs Shorts-style format, duration, and viewing-target guidance checked on 2026-07-09.
- [references/youtube/youtube-resolve-delivery.md](references/youtube/youtube-resolve-delivery.md): DaVinci Resolve delivery workflow for YouTube upload files.
- [references/davinci/tools-davinci-resolve.md](references/davinci/tools-davinci-resolve.md): DaVinci Resolve Studio, Python API automation, project creation, timeline fps, and assembly notes.
- [references/ffmpeg/tools-ffmpeg.md](references/ffmpeg/tools-ffmpeg.md): FFmpeg-family command reference for
ffprobe,ffmpeg, and related media-preparation commands. - [references/zundamon/tools-zundamon.md](references/zundamon/tools-zundamon.md): Zundamon narration text preparation and VOICEVOX Engine HTTP API notes.
- [references/troubleshooting/quality-check.md](references/troubleshooting/quality-check.md): verification checklist for generated media, including VLC playback checks.
- [references/VERSION.md](references/VERSION.md): source of truth for this skill version.
Templates
- [templates/response-template.md](templates/response-template.md): concise completion report structure.
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: igapyon
- Source: igapyon/miku-media-proc-skills
- License: Apache-2.0
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.