choplin/explainer-studio/explainer-audio-guide

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Coordinate creation, review, VOICEVOX preparation, narration, continuation, or selective rebuilding of one audio guide from an exact existing source or report. Use as the user-facing audio entry point; delegate all content and synthesis work to the audio phase-owner skills.

Install

npx vskill@latest install choplin/explainer-studio/explainer-audio-guide

SKILL.md

Coordinate one audio guide

Own orchestration only. Resolve the requested terminal Artifact, reconcile exact inputs and existing outputs, compute the smallest dependency closure, and invoke the audio phase owners. Do not select content, write dialogue, review it, convert kana, or synthesize audio inside this skill.

Apply [[explainer-content-workflow-base]] and read its coordination and Artifact contracts when a profile coordinator delegates a planned audio slice. For an unrelated exact Markdown source or report, use the standalone contracts of the same phase owners. If source extraction, report generation, Content modeling, site generation, or cross-medium consistency is required, route the request to [[book-explainer]] or [[paper-explainer]] instead of expanding this coordinator.

Resolve the invocation

Accept one exact starting point:

  • a standalone Markdown source or report;
  • a Guide plan and its exact source;
  • a readable dialogue with its Guide plan and source;
  • a dialogue review with all inputs it identifies;
  • a validated VOICEVOX dialogue; or
  • a workflow run request, Content brief, and Run Manifest delegated by a profile

coordinator.

Resolve the desired terminal state from the request: Guide plan, reviewed readable dialogue, validated VOICEVOX dialogue, or narrated audio. Ask only when that terminal state, content policy, output replacement, or optional human review is materially ambiguous. Never ask the user to select internal phases.

For standalone creation from Markdown, resolve the Guide policy once, including guide kind, listener destination, coverage, depth, audience, supporting detail, pacing, segmentation, duration treatment, and cast pair. Pass those exact choices forward; later owners must not reconstruct them from conversation.

Reconcile and plan

Inventory the starting Artifact and every conventional or Manifest-planned downstream path. Validate recorded provenance and SHA-256 digests. Classify each relevant output as reuse, create, replace, or untouched, then compute the prerequisite closure from the earliest missing, invalid, or intentionally changed Artifact.

In workflow mode, treat the immutable Run Manifest as execution authority. Ask [[explainer-run-plan]] for a new Manifest when the terminal set, policy, input digest, or replacement closure changes. Accept an existing Manifest only for an explicit exact-run resume. In standalone mode, do not silently overwrite any existing output; obtain one explicit replacement decision covering the named affected paths.

Never reuse a downstream Artifact when any recorded upstream digest differs. Show the compact reuse / create / replace / untouched slice before a destructive replacement or when the valid starting point is ambiguous.

Execute the dependency closure

Invoke only the owners required for the resolved terminal Artifact, in order:

  1. [[explainer-audio-guide-plan]] creates or replaces the Guide plan.
  2. [[explainer-audio-dialogue]] writes the human-readable dialogue.
  3. [[explainer-audio-dialogue-review]] independently reviews that exact digest.
  4. After result: pass and any selected human gate, [[explainer-audio-voicevox]]

writes and validates the propositionally paired synthesis script.

  1. [[explainer-audio-narrate]] synthesizes audio when audio is the terminal.

Pass exact paths and digests returned by one owner to the next. A request for a readable dialogue includes independent review; do not present an unreviewed draft as a successful terminal Artifact. A request for VOICEVOX or audio always includes the same review prerequisite.

Route review and runtime failures

On result: revise, stop forward execution and obey the review's return_to:

  • guide-plan for a defective destination, coverage assignment, required

facets, Guide structure selection, or explanatory spine;

  • dialogue when the Guide plan is sound but its realization, fidelity,

conversational moves, or spoken orientation fails.

In workflow mode, create a new Manifest that replaces the returned-to Artifact and every requested dependent output. In standalone mode, show those exact replacement paths and obtain replacement authorization before rerunning. Never patch the review into pass, reuse it for another dialogue digest, or continue to VOICEVOX while revision is unresolved.

Keep conversion and synthesis failures local when upstream digests remain valid: kana or pairing failures return to [[explainer-audio-voicevox]]; engine, encoding, or output failures return to [[explainer-audio-narrate]].

For a selected workflow dialogue-pilot gate, present the exact passing readable dialogue and review, persist the user's decision through the shared checkpoint contract, and continue only from approved. For standalone human inspection, use the reviewed dialogue as the current terminal and resume later from its exact paths and digests.

Finish

Report the starting inputs, resolved terminal, produced and reused Artifacts with digests, review and validation results, checkpoint decisions, replacements, and unresolved blockers. Do not claim success beyond the last validated Artifact.

Success criteria

  • [ ] One exact starting point and terminal Artifact determine the execution.
  • [ ] Only the smallest valid dependency closure ran.
  • [ ] Every owner received exact paths, digests, policy, and output authority.
  • [ ] Reviewed dialogue is the minimum successful content terminal.
  • [ ] No revise result or stale digest flowed into VOICEVOX or narration.
  • [ ] Failures returned to their owning phase without rerunning valid upstream

work.

  • [ ] The final handoff distinguishes produced, reused, replaced, and blocked

Artifacts.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/choplin/explainer-studio/explainer-audio-guide/badge)](https://verified-skill.com/skills/choplin/explainer-studio/explainer-audio-guide)

Something wrong with this skill or its page? Tell us.