Appendix C — Templates

Status: DRAFT 1 (2026-09-21), per DESIGN v6 §12. One template per document the book asks the learner to write, in chapter order. Copy, fill, keep in the film folder. Templates marked (from PRODUCTION.md) follow the reference production's own form; the rest are the book's.

C.1 Working agreement (Chapter 0)

DIRECTOR: <name>          AGENT(S): <names>          SECOND READER: <name>          FILM: <title>
Decisions only the Director makes: story, casting, which take, cut/rewrite, lock, delivery.
Spend rule: every paid job needs a fence and a "go" from the Director in chat. One job per go.
Ledger rule: one row per job, read, decision, fence, correction; append only; numbers only after the read.
Report format: WHAT I READ / REMAINING ISSUES / NEXT ACTION.
Take rule: the Director chooses with the five-step protocol; the agent's measurements are read after the sentence is written.
Never (from Chapter 1 on): delete or overwrite. Never (from Chapter 2 on): <the destination's rights rules, copied here>.
Clicks the agent hands back to the Director: <binding submissions, unpriced buttons, anything irreversible>.

C.2 Ledger row (Chapter 0)

timestamp, kind (job|read|decision|fence|correction), subject, detail, price_or_credits, status

Conventions: ISO timestamp with zone; detail quotes the source of every number ("verification.json: 240/240"); a correction row names the row it corrects and never edits it; a job row is written twice, once when queued (status queued, price read at the button) and once when read (status read, actual charge from the account).

C.3 Fence (Chapter 0, used from Chapter 2)

FENCE — <job name>, <date>
Purpose: <the beat this serves; the card's goal line>
Inputs: <references by file name, in the order they will be attached; none from outside the project>
Settings: <model, resolution, duration, sound on/off, count>
Expected price: <read at the button before the click>
Check afterwards: <the strip, sheet or transcript that decides pass/fail; who judges what>
Budget line: <card tier and takes used so far / budget>

C.4 Review report (Chapter 0)

WHAT I READ      — the files, strips, measurements actually looked at, with paths
REMAINING ISSUES — numbered, each with the evidence, none with a verdict the evidence does not support
NEXT ACTION      — one concrete step, its cost if any, waiting on which approval

C.5 Take-comparison sheet (Chapter 0, used from Chapter 11)

CARD: <id>    GOAL LINE: <from the planning card>
Takes: A <file>   B <file>
Watched full speed, sound on: A then B   [ ]
Watched silent: A then B                 [ ]
Sentence (which, and why, against the goal line):
<one sentence naming something visible or audible>
Agent measurements read AFTER the sentence: <paths>
Do they contradict the sentence? <no / yes: what; watched again: result>
Circle take: <A / B / neither → retake or rung>
Turn shot only (Chapter 11 step 7): reader's blind pick <A / B>, initials; agrees with mine <yes / no: one sentence, by the goal line>

C.6 Learner case (four fields; Chapter 2 on)

CASE <chapter>-<letter> — <one-line title>
What I saw (with file): 
Why I think so: 
What I changed: 
Would this happen on your tool? 

C.7 Instructor case (six fields; Appendix E and cases with established attribution)

Observation (what, where, file):
Attribution (with evidence):
Controlled comparison run: yes / no (what was held constant):
Remedy applied:
Residual uncertainty:
Transfer limit: tool-specific / HALFWAY-specific / general

C.8 Director's-notes line (Chapter 8 on)

<shot> | should feel: <one phrase> | feels: <one phrase> | keep / retake (rung n) / cut

C.9 Beat sheet row (Chapter 3)

# | Beat title | Whose (owner) | What changes (two sentences) | Value before → after | Location | Run s | Lines | Cards

Gate check per row: the value column changes; the owner is one character; the location exists in the bible.

C.10 Coverage intent, one per scene (Chapter 5)

SCENE <n> — <title>
Whose scene: <character>
Where the turn is: <the beat, in one line; which card carries it>
What the audience knows that a character does not: 
The exit: <the last image of the scene, and what it hands to the next>
The line: <the axis between the two characters; which side the camera stays on>
Reaction cards: <one per dialogue beat, listed>

C.11 Spatial map, one per location (Chapter 4, extended in Chapter 5)

LOCATION <asset id> — <name>
Fixed objects and their positions in words (door, window, bed, table, lamp), as seen from the camera side of the line:
Character A enters from / sits at / faces:
Character B enters from / sits at / faces:
Eyelines: A looks <screen left/right> to B; B looks <screen right/left> to A
Light source and direction:
What must never appear (a second door, a mirror, a window where the wall is):

C.12 Planning card (Chapter 5) and final prompt (Chapter 11)

Planning card:

CARD <scene><letter>   GOAL LINE: <what this shot is for, one line>
Location asset: <id>       Characters and state: <who, wardrobe, prop state, emotional state>
Shot size (edge to edge): <what fills the frame>     Eyeline: <stated>
Running time: <s>   Tier: turn / dialogue single / insert   Take budget: 3 / 2 / 1 (from the tier)   Takes used: <n>
Replacement card (rung 2–3)? <no / yes: replaces <id>, own budget from its tier>
Line (at most one speaker on camera): <text> — intention line: <want / obstacle>
Reaction card paired with: <id or none>
References budget: <count>   Start frame: <frame-zero still id or none>

Final prompt (written from the card; sent to the tool):

[ordered references, identity first, then location, then prop, then blocking-hands reference last]
<shot size edge to edge>. <who, state, wardrobe>. <the motion beat, written as one continuous action with its floor contact and furniture distance>. <eyeline>. <the line, if any, and the affect of the face while saying it>. <what the frame must not contain>.

C.13 Intention line (Chapter 8)

LINE: <text>   SPEAKER: <name>
Wants (right now, from the other person): 
In the way: 
The listening beat before it: <what the other face does>
Accept by ear when: <one audible thing that shows the want>
Blind pair (Chapter 8 step 5): my pick <first / retake>, written <date> before the listening
Reader's pick: <first / retake / could not tell>   Reader's answer to "what does it want": <their words, initials>

C.14 Animatic viewing form (Chapter 10, second reader)

Watched cold, once, sound on, no pauses   [ ]   Date:        Initials:
Retell the turn in two sentences:
The slowest minute (mm:ss – mm:ss):
Who is who (name the two people as you understood them):
One thing you did not understand:

C.15 Build manifest and verification (Chapters 1, 11–15)

manifest.json: cut name, fps, planned duration, one item per shot (name, source, start, duration, frames, timeline start, segment path). verification.json: full_decode, expected_frames, actual_frames, duration, status ("Technical decode only. Not a quality approval."). Both written by templates/build_cut.py; a cut is a new folder every time.

C.16 Delivery sheet (Chapter 2)

DESTINATION: <name>   RULES PAGE URL:            READ ON: <date>   LAST UPDATED (as the page says):
Format: <aspect>   Resolution:   Frame rate:   Container/codec:
Runtime: floor    ceiling    margin we keep:
Loudness target:            True-peak ceiling:
Watermark / packshot / title card required:
Subtitles required: <language, burned-in or file>
Rights rules (copied verbatim, each with its page section):
  real likeness / voice:
  non-AI or licensed music:
  assets generated off-platform:
  other:
Provenance: <where every asset and audio file must live for the verifier>
Social post required: <yes/no; where; what>
Submission: <how; is there a receipt; can a submitted file be replaced, until when>

C.17 Cutoff timeline (Chapter 2)

| Event | Date and time (with zone) | Source (page, read on) | Our plan date | Margin |
| generation window ends | | | | |
| submission cutoff | | | | |
| replacement allowed until | | | | |
| social post due | | | | |
| rules page re-read | (submission day) | | | |

C.18 Screening form (Chapter 16, one per viewer)

Viewer:            Date:            Watched once, cold, sound on, no pauses   [ ]
1. What happened at the turn? (two sentences)
2. Who are the two people? (one line each)
3. Any line where the mouth and the voice did not match? (mm:ss)
4. What did you feel at the end? (one word or one sentence)
5. Did either person's face ever look like a different person? (yes/no; if yes, roughly when)
6. Did the sound ever draw your attention to itself (music over a voice, a jump at a cut, a breath)? (yes/no; if yes, when)

C.19 Capstone rubric (Chapter 16)

Four rows, four levels, scored from the three forms by count so two people tallying get the same level. Q1 = turn retold correctly (the learner judges "correctly" against the beat sheet and writes why); Q2 = both people named; Q3 = lines named; Q5 = face note; Q6 = sound note.

1 2 3 4
Story Q1 correct: 0 of 3 1 of 3 2 of 3 3 of 3
Identity Q2 both named: fewer than 3; or any Q5 yes from 2+ viewers Q2 3 of 3 and Q5 yes from 1 viewer Q2 3 of 3, Q5 all no, but one viewer hesitated on a name Q2 3 of 3, Q5 all no, no hesitation
Sync Q3: 2+ distinct lines named, or one line named by 3 one line named by 2 one line named by 1 no line named
Sound Q6 yes from 2+ viewers Q6 yes from 1 viewer naming a voice covered by music Q6 yes from 1 viewer naming a cut or a breath Q6 all no

Tie rule: when a form fits two levels, the lower one counts.

Self-critique (required, one page): the one shot you would reshoot, which rung you would use, and what in the bible or the card would change first.