tk-video

Explain Through Video

Produce the selected MP4 or GIF from observed software states or explicitly labeled explanatory cards. Existing-video editing and unrelated filmmaking are outside this skill. A movie communicates a flow; it does not establish application correctness, latency, performance or reader understanding.

Plan and collect

Resolve the audience, claim, selected format and available scene sources. Prefer stills when motion itself is not the claim. A short flow around ten seconds often fits GIF; longer explanation fits MP4. Reuse supplied material and approved choices without another routine question. For a human-facing tutorial, resolve narration and cursor/input overlays with format in the same existing choice, only where not already selected. Give captions reading time and measured narration at least 800ms of breathing room. Plan each crop and highlight its region before zooming in.

List each scene's source, caption, optional narration and hold or observed timestamp. Label rendered cards with a corner source tag in the audience's language and declare source_in_picture: true in each tagged frame; otherwise retain a localized caption label. Never reenact unavailable software behavior or present a diagram as a capture. Keep claims, source versions and unverified intervals explicit. For real browser states, invoke tk-browser-verify with phase: media-capture for a demo or explanation without acceptance criteria; an explicitly requested acceptance check retains its verification contract and optional motion-evidence supplement. For native windows invoke tk-app-verify. Those owners retain provider, authentication, test data, actual input/capture and resource lifecycle. Return to this skill with ordered, publication-safe frame paths and provenance. Do not import a sibling skill's script path. If that owner is unavailable, report the missing capture; do not invent a replacement.

Create cards with an already available renderer, retaining their authored source. HTML cards may use an already approved headless rendering boundary; installed plotting/rendering tools are also eligible. No dependency installation or user-profile attachment follows from choosing video.

Assemble and inspect

Read assembly for the scene/capture manifest, owned paths and command. Use explainer holds for presentation and preserve actual timestamps in evidence mode. Keep assertion screenshots unchanged and independent from the movie. Use H.264 MP4 at the capture's frame rate capped at 30fps, with optional mono 48kbps AAC; GIF and explainer stills default to 10fps. MP4 defaults to CRF 18; use the assembler's quality/rate overrides and preserve source resolution. Follow prerequisite recovery for unavailable tools. If hard-caption support is missing after installed-tool discovery, present the tk-wizard installation-guidance question before returning the blocked result. Do not implement a frame-compositing workaround.

For selected generated narration, invoke tk-tts through supported skill loading/handoff; read narration for measured audio-to-scene alignment. Reuse supplied audio when selected. Assembly consumes ordinary WAV files and imports no TTS engine. Selecting TTS requires actual accepted non-silent audio; missing synthesis is not a voiced silent preview.

Run the finished-media checks and inspect the generated contact sheet plus representative decoded frames, including the initial/final state and every material caption or scene transition. Return human_action: listen with the names/numbers/terms to check and relevant sample paths for selected narration; retain it as pending until the user confirms meaning, pronunciation and clipping. A decode pass is not visual or pronunciation inspection. Distinguish uninspected intervals and report real limitations. Repair a defective cut and rerun affected checks; never edit the receipt to make the movie pass.

Return

Return ready media with its absolute path, mode/timing, sources, caption/audio state, media checks, inspection limits and publication selection. Report blocked or failed quality separately from application verification. Missing hard-caption support retains the soft-caption fallback, not a claim of completed burned captions. Preserve inputs and failure evidence; clean only known owned transients only after the user confirms the selected result and its decode/inspection completes. Until then preserve reusable frames/audio/manifests and report retained frame bytes. Release capture resources through their owning verifier.

Generation authorizes no publication. Hand selected attachments to tk-pr-media only when upload is separately authorized. This skill changes no application source, Git state, global settings, credentials. It creates only the requested movie, scene sources and run-owned evidence, plus the selected audio artifacts. tk-tts owns its separately approved installation and model cache.

Design sources and deliberate omissions are recorded in sources.

Narration Choice

Ask the creator to choose TTS, supplied audio or no narration before any video or spoken-slide production. Reuse an answered parent/user choice; video or a voice preference is not TTS consent. Show current tk-tts model/voice/device/setup options before synthesis/assembly. Preserve no narration; offer MP4 for GIF sound. Static HTML/text needs no unrelated voice interview. Installation/download/publication require their own authority; playback mute/fallback are separate.

Artifact Paths

Artifacts require task authority. Before writes/checkout/transport/ignore setup, read artifact paths for Git exclusion/path/ownership checks. Reuse identical in-context content; hash-check uncertainty and reread changes/loss; always recheck safety. Use .tigerkit/ unless a destination is explicit. Chat-only: skip file/ignore setup and this reference. This grants no unrelated mutation/publication authority. Prefer fresh directories and supported trash for authorized non-secret cleanup; target only recorded literal resolved absolute run-owned paths, never cd plus glob/relative deletion.

User Questions

Investigate facts; reuse authorization. For user-owned decisions read question rounds; otherwise proceed. Plain chat, even one approval: standalone ---, then ❓ **Q<n> · <title>**: <question> and ➡️ <recommendation> per item. Ask only settled-prerequisite questions; separate approval. Questions come last; put status, risks and plans before them.

Skill Feedback

For a reusable incident, retain minimal non-secret evidence in the current conversation and suggest tk-retro; do not write it to persistent memory unless the user asks. Invocation, feedback artifacts, installed-skill edits and publication require separate authorization through the owning workflow; this pointer grants none.

Reader-facing Result Summary

Return a brief result with the outcome, actual verification, material limitations and direct deliverable links. Distinguish plans/skipped checks from executed work; retain outstanding manual checks and real cleanup, blocker or human-action details. Never claim unverified completion or cleanup. Preserve the owner's required artifact, publication preview, approval boundary, machine return, exact status/footer and final question block; keep nested returns compact and omit empty sections.

Output Notation

Use ASCII numbering (1. Item, (1) Item) and - Item, not enclosed/keycap numerals. Show complete absolute clickable file paths, including examples, without ellipses. Preserve exact code, commands, identifiers, URLs, quotations and verified UI labels unless their edit is authorized.

For chat comparisons with multiple shared attributes, use narrow GFM pipe tables (about four columns, header, short cells); otherwise use lists. No box-drawing, fenced or empty tables. Owner-specific previews take precedence; publication bodies and file artifacts retain their own table formats.

In rendered Markdown prose, use hyphens for ranges and escape literal tildes as \~; preserve exact code, quotations and identifiers.

For authorized temporary input files, follow Artifact Paths before creation and consumption; a prefilled template is not completed input. This grants no write authority.