Post-Confirmation Production

This skill is only responsible for safely delivering already-written production specifications to the adapter configured in the runtime environment. Image prompts remain the responsibility of

$short-drama-image-prompts

, video prompts belong to

$short-drama-video-prompts

, dialogue and recording sheets belong to

$short-drama-write

, and voice identity belongs to

$short-drama-assets

Quick Start

Only after the creator explicitly requests actual generation, take the current prompt from

图片提示词.md

分镜.md

, or

视频提示词.md

, and create a bounded run job. A creator-first job's

source

must point to the current Markdown that owns this prompt, and

source_entry

must name the document's allowed secondary heading:

图片提示词.md

uses

IMG-*

视频提示词.md

uses

MOTION-*

分镜.md

uses

SHOT-*

(modality is

image

, body is taken from that shot's

### 冻结关键帧提示词

). This selector only extracts the first frame; the tail frame must first be established as a separate

IMG-*

state-graph entry by the image prompt stage, and then an image job is prepared from

图片提示词.md

. Creators may also provide an existing start frame—generating the image in their own tool and placing it into the project is equally valid; if images never enter the project, use

PLAN-...

from the storyboard; after the output lands in

剧集/<EP>/制作成果/images/

, the storyboard owner can bind it as a

REF-...

with

用途：起始帧

. When real reference images exist, you must also fill in each slot of

reference_bindings

with order, path, Chinese name, purpose, and the allowed/forbidden control scope; the built-in video adapter reads the

用途

field here as the vendor's own role and only accepts its published values (MiniMax uses

first_frame

last_frame

reference_image

reference_video

reference_audio

, Seedance uses three

reference_*

values); a job with reference images but no bindings will fail directly—it will not guess a role. Local images are sent by the built-in adapter as base64 data URIs directly; no self-hosted upload service is needed. When an entry's "input reference image" is written as

PLAN-...

, those images live in the creator's own tools and there is no file in the project to send:

prepare

fails directly and explains that real files must first be placed in the project and rewritten as

REF-...

; it does not use the plan as input and does not silently drop references.

references

can be omitted and generated from binding order, or provided as an explicit mirror in the same order. Output goes to

剧集/<EP>/制作成果/

; this job is a temporary input for the production tool, not a sixth creation document:

bash

python3 {skill_directory}/scripts/production_tool.py prepare <project> --job <temporary-job.json>

First show the complete preview from

prepare

; no vendor is called at this point.

Hard Gate

Every production must go through the following four steps, in order, and they cannot be merged:

Create a bounded job: one modality, explicit quantity, complete prompt/spec, reference files, parameters, output path, and adapter profile.
Run
```
prepare
```
and show the returned complete preview to the creator, especially quantity, prompt, source entry, reference bindings, references, outputs, overwrite, and adapter. A creator-first job mechanically checks at this step the copyable prompt in the selected heading, as well as reference image slots, order, path, Chinese name, and control boundaries; any drift fails closed.
Wait for the creator to explicitly confirm after seeing this preview. Only run
```
confirm
```
when the creator explicitly agrees to this current task; "continue," "do them all," "budget is fine," upstream content being accepted, or having previously confirmed another version do not count as confirmation for this production.
Run
```
run
```
. It consumes one confirmation before starting the adapter; after either success or failure, any further execution requires re-confirmation, to prevent failed retries from accidentally generating a second charge.

If the job, prompt, parameters, output path, or direct input change in any way, the old confirmation immediately becomes invalid. You must not fill in confirmation on behalf of the creator. The currently confirmed job is the only work unit for this round; after execution completes, report the result and return control, without automatically preparing the next batch or starting review.

The "input reference image" path in

分镜.md

is only a readable basis and usage intent during the creation stage, not a production-input snapshot. When entering production, a creator-first job must establish bindings from the corresponding entry in

图片提示词.md

视频提示词.md

; the

reference_bindings

references

, and the confirmed job shown by

prepare

are the authority for which file bytes the adapter actually reads this time and what each is allowed to influence. Non-creator structured specs may omit

source_entry

reference_bindings

and continue using only explicit

references

; but any new image/video job whose

source

points to canonical

图片提示词.md

视频提示词.md

is forced to use the corresponding selector and cannot downgrade by omitting fields. Old jobs that were prepared and written to disk before the upgrade can still be read by their original fingerprint. New production results are not automatically backfilled or refreshed into the storyboard; if they need to become subsequent input, the storyboard owner revises the document, then creates a new job and re-previews and re-confirms it.

Commands

Only after entering the production boundary, write the current prompt and run parameters as a temporary JSON; for video and image jobs, the

parameters.prompt_language

follows the prompt language already parsed from the current copyable body, so that adapter-appended reference constraints use the same language instead of falling back to fixed English. Do not pre-create jobs for every prompt during the creation stage. Format and adapter contract are in adapter-contract.md. Commands are provided by production_tool.py, then run:

text

python3 <this-skill-directory>/scripts/production_tool.py prepare <project> --job <job.json>
python3 <this-skill-directory>/scripts/production_tool.py confirm <project> --job-id <id> --confirmation "CONFIRM <id> <code>"
python3 <this-skill-directory>/scripts/production_tool.py run <project> --job-id <id> --adapter-config <outside-project-config.json>
python3 <this-skill-directory>/scripts/production_tool.py status <project> --job-id <id>
python3 <this-skill-directory>/scripts/production_tool.py collect <project> --job-id <id> --adapter-config <outside-project-config.json>
python3 <this-skill-directory>/scripts/production_tool.py audit <project>

prepare

only validates and previews; it does not produce.

confirm

only saves a one-time confirmation bound to the current job fingerprint.

run

is what starts the adapter.

collect

only retrieves the result of a task that was already submitted and already billed; it does not resubmit and does not need new confirmation—see "Interruption Does Not Mean Re-running" below.

audit

only reconciles the local task history, post-failure recovery, duplicate content attempts, and current output bytes; it does not call vendors and does not write technical success, file existence, or hash consistency as media-quality conclusions. While an unresolved

running

attempt exists for the same job, re-running prepare, confirm, or run is forbidden; first wait for completion or investigate the leftover attempt.

Input Selection

image: read the current
```
IMG-*
```
copyable body from
```
图片提示词.md
```
, or the current
```
SHOT-*
```
from
```
分镜.md
```
frozen keyframe body, plus required reference images and explicit output dimensions/quantity; creator-first jobs use
```
source_entry
```
to lock this entry. Asset board uses
```
IMG-*
```
; a particular shot's start image uses
```
SHOT-*
```
; they do not replace each other.
video: read the current
```
MOTION-*
```
copyable body from
```
视频提示词.md
```
, and cross-check the corresponding shot, frozen keyframe, duration, and aspect ratio in
```
分镜.md
```
; creator-first jobs use
```
source_entry
```
to lock this entry. When a continuous segment continues from the previous segment's generation result, the next segment's job binds both the previous segment's actual video and the actual tail frame obtained from that video, while retaining the distinct responsibilities of
```
continuity_video
```
and
```
actual_tail_frame
```
; vendor roles are translated by the target-model adapter. This group of H3 inputs is uniformly translated to
```
reference_video + reference_image
```
, not mixed into
```
reference_video + first_frame
```
; do not substitute the planned tail frame or text description for the real file.
tts: read the original sentence and performance requirements from
```
剧本.md
```
; voice reference is explicitly provided by the user or existing media. Do not modify lyrics in production jobs, and do not create a sixth creation document for TTS.
music: read the creator-confirmed timeline music section from
```
视频提示词.md
```
; theme songs use confirmed lyrics, pure BGM carries no lyrics. When the vendor cannot precisely promise duration, after generating the source audio track,
```
$short-drama-edit
```
still handles placement, looping, fades, and dialogue ducking according to the mixing intent in the document.

A single job does not mix modalities. Large batches are split into small jobs where the creator can see clear quantity and cost boundaries; do not implicitly stuff an entire season into one confirmation for convenience.

Adapter Boundary

Adapter configuration must live outside the project and contain only the argv command and timeout; credentials are read by the adapter itself from the process environment or system credential store. Project jobs, confirmation records, run logs, and Dashboard must not store secrets.

The script invokes the argv array via JSON stdin, does not use a shell, and does not concatenate commands. The adapter returns local temporary files; the tool only accepts results that exactly match the confirmed targets, and atomically copies complete files into the project

剧集/<EP>/制作成果/

directory. The project and upstream Skills do not hardcode vendors, models, or soon-to-change APIs.

The built-in image/video compiler appends a deterministic reference-semantics note to the vendor prompt (Chinese name, purpose, allowed control, and forbidden control scope) in order, based on the confirmed

reference_bindings

; it does not mistake slot names for text to be rendered into the frame. External adapters must also preserve this semantics or explicitly refuse, and must not silently lose control boundaries by merely uploading files.

This skill optionally provides four stdlib adapters, all selected via an out-of-project adapter config, with credentials read only from the runtime environment:

Seedance: model/Endpoint ID must be explicitly configured by the account; compiler supports official image, video, and audio reference roles; the built-in runtime still rejects local reference files when no trusted upload is configured.
GPT Image 2: no reference image → generation; with reference image → edit; fixed high-fidelity reference, validates size, format, and transparent-background constraints.
MiniMax Music: uses
```
music-3.0
```
and hex results; distinguishes theme song and pure BGM; does not fabricate duration request fields.
MiniMax Speech:
```
tts
```
modality; both model and
```
voice_id
```
must be explicitly given by the account and documentation; no built-in voice list. Synthesizes only from preset voices; no voice cloning—cloning is a licensing issue. Creator-authorized reference recordings go through
```
输入/
```
as ordinary creator input.
MiniMax H3 Video: model ID, resolution set, and duration range must be explicitly configured by the account; prompts go into the text item of
```
content
```
, reference images are bound by explicit role, and local references fail closed without a trusted upload. This model generates sound in the same pass as visuals; the writing impact is described in the video prompt skill's target-model capability profile.

These adapters are validated request contracts, not guarantees of account availability or generation quality; formal production must still pass the above per-confirmation gate, and a review Skill judges the quality of the output.

The repository's

fixture_adapter.py

is for offline testing only and does not represent real generation quality or the default production adapter.

Interruption Does Not Mean Re-running

A video task is billed at the moment of submission, not when the result is received. Everything after submission—polling for minutes, downloading— can be interrupted by a killed process, network loss, or machine sleep. The built-in adapter therefore writes the vendor task ID into

handle_path

as soon as it receives it (before the first poll), and this path is not deleted with this attempt.

So after an interruption, there are three deterministic actions that avoid spending again:

```
audit
```
reports the unfinished attempt with its task ID as
```
orphaned_provider_job
```
, with
```
action
```
set to
```
collect_before_retry
```
;
```
collect
```
uses that ID to retrieve the result and marks this attempt as successful;
Only if
```
collect
```
also confirms that the provider really failed do you go down the old path of re-confirming and resubmitting.

Do not resubmit directly when
audit
reports
orphaned_provider_job
—that would be paying a second time for the same shot. It is intentional that

collect

does not go through the confirmation gate: the gate prevents accidental spending, and collect does not spend; if it also required re-confirmation, the cheapest path after interruption would become paying again, exactly what the gate is meant to prevent.

Results and Review

On success, report the actual output path, media type, and run status; do not write "adapter returned success" as a quality conclusion. After multiple tasks or retries, first run

audit

: terminal failures route via

retryable

; retries still require new explicit confirmation; if output is missing or a file's hash or size no longer equals the run record, first recheck the current bytes or re-produce.

repeated_content

is only a cost and diagnostic signal; it cannot automatically determine whether the same-text retry is reasonable or unreasonable;

running_attempt

is a pending operational status, and audit must return attention. Failures are routed three ways. Technical failures such as timeout, throttling, and server errors can be retried with a cap. When the failure message names which input was rejected— prompt text, reference image, or audio—fix that item before resubmitting: if text was rejected, rewrite that sentence, changing "fist slams into the other's face, blood drips along the chin" to "fist swings empty, the other sidesteps, the cup on the table is knocked over"; if the reference image was rejected, replace it with an image whose composition and character are consistent and whose content is compliant; if audio was rejected, re-record that line. Write the change into a new job and re-run prepare, letting the creator see in the preview which item changed, then confirm; resubmitting the same content without change produces no fix, only a charge. Duplicate-content defects return to the corresponding prompt/spec owner. Production does not automatically enter editing. When the material needs to be assembled into a final cut, the user names

$short-drama-edit

; it only selects among existing frames, and does not retroactively change jobs from this stage or generate new material. If quality review is needed, the report can hand existing results to

$short-drama-review

separately; do not auto-start review from within the production call. The Dashboard is only for displaying these files and run summaries; it does not provide adapter settings or production buttons.

Installation and Maintenance

Run offline self-tests only during installation, upgrade, or troubleshooting; not during normal creation or production preparation:

bash

python3 scripts/selftest.py
python3 scripts/provider_adapters.py --selftest

short-drama-produce

NPX Install

Tags

SKILL.md Content (Chinese)

Post-Confirmation Production

Quick Start

Hard Gate

Commands

Input Selection

Adapter Boundary

Interruption Does Not Mean Re-running

Results and Review

Installation and Maintenance