@cliangdev/conductor 0.31.1 → 0.32.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/assets/claude/agents/conductor-content-art-director.md +30 -7
- package/assets/claude/agents/conductor-content-creative-director.md +7 -2
- package/assets/claude/agents/conductor-content-ideator.md +9 -3
- package/assets/claude/agents/conductor-content-scriptwriter.md +12 -4
- package/assets/claude/commands/conductor/content.md +3 -1
- package/assets/claude/commands/conductor/creative.md +14 -8
- package/assets/claude/skills/conductor-content-studio/SKILL.md +47 -17
- package/assets/claude/skills/conductor-content-studio/references/capabilities.md +11 -3
- package/assets/claude/skills/conductor-content-studio/references/design-critique.md +52 -0
- package/assets/claude/skills/conductor-content-studio/references/idea-template.md +3 -3
- package/assets/claude/skills/conductor-content-studio/references/layout-guide.md +52 -0
- package/assets/claude/skills/conductor-content-studio/references/publish-pack-template.md +8 -3
- package/assets/claude/skills/conductor-content-studio/references/rubric.md +6 -0
- package/assets/claude/skills/conductor-content-studio/references/slideshow-formats.md +121 -0
- package/assets/claude/skills/conductor-creative/SKILL.md +42 -8
- package/assets/claude/skills/conductor-creative/references/brand-kit-setup.md +3 -1
- package/assets/claude/skills/conductor-creative/references/design-critique.md +52 -0
- package/assets/claude/skills/conductor-creative/references/hooks.md +47 -0
- package/assets/claude/skills/conductor-creative/references/layout-guide.md +52 -0
- package/assets/claude/skills/conductor-creative/references/slideshow-formats.md +121 -0
- package/assets/claude/skills/conductor-publisher/SKILL.md +2 -2
- package/assets/cli-manifest.json +5 -0
- package/assets/fonts/OFL.txt +93 -0
- package/assets/fonts/tiktok-sans-600.ttf +0 -0
- package/assets/fonts/tiktok-sans-700.ttf +0 -0
- package/assets/fonts/tiktok-sans-800.ttf +0 -0
- package/dist/lib/plugin-assets.d.ts.map +1 -1
- package/dist/lib/plugin-assets.js +7 -0
- package/dist/lib/plugin-assets.js.map +1 -1
- package/dist/lib/slideshow/compose.d.ts +37 -0
- package/dist/lib/slideshow/compose.d.ts.map +1 -0
- package/dist/lib/slideshow/compose.js +160 -0
- package/dist/lib/slideshow/compose.js.map +1 -0
- package/dist/lib/slideshow/css.d.ts +3 -0
- package/dist/lib/slideshow/css.d.ts.map +1 -0
- package/dist/lib/slideshow/css.js +103 -0
- package/dist/lib/slideshow/css.js.map +1 -0
- package/dist/lib/slideshow/html.d.ts +20 -0
- package/dist/lib/slideshow/html.d.ts.map +1 -0
- package/dist/lib/slideshow/html.js +192 -0
- package/dist/lib/slideshow/html.js.map +1 -0
- package/dist/lib/slideshow/index.d.ts +4 -0
- package/dist/lib/slideshow/index.d.ts.map +1 -0
- package/dist/lib/slideshow/index.js +4 -0
- package/dist/lib/slideshow/index.js.map +1 -0
- package/dist/lib/slideshow/spec.d.ts +109 -0
- package/dist/lib/slideshow/spec.d.ts.map +1 -0
- package/dist/lib/slideshow/spec.js +167 -0
- package/dist/lib/slideshow/spec.js.map +1 -0
- package/dist/mcp/index.d.ts +403 -0
- package/dist/mcp/index.d.ts.map +1 -1
- package/dist/mcp/index.js +21 -0
- package/dist/mcp/index.js.map +1 -1
- package/dist/mcp/tools/slideshow.d.ts +42 -0
- package/dist/mcp/tools/slideshow.d.ts.map +1 -0
- package/dist/mcp/tools/slideshow.js +77 -0
- package/dist/mcp/tools/slideshow.js.map +1 -0
- package/package.json +1 -1
|
@@ -13,9 +13,10 @@ send to the draft renderer, so it has to be valid on the first try.
|
|
|
13
13
|
## Boundaries
|
|
14
14
|
|
|
15
15
|
- Read only the files you are given: `script.md`, `context/media.json`, `context/brand-kit.json`,
|
|
16
|
-
`references/shotlist-template.md`, `references/platform-specs.md`, `references/capabilities.md
|
|
17
|
-
`references/publish-pack-template.md` when given.
|
|
18
|
-
- Write only your output files: `direction.md` and, for `RENDER`, `creative-spec.json
|
|
16
|
+
`references/shotlist-template.md`, `references/platform-specs.md`, `references/capabilities.md`,
|
|
17
|
+
`references/layout-guide.md`, `references/design-critique.md`, `references/slideshow-formats.md` and `references/publish-pack-template.md` when given.
|
|
18
|
+
- Write only your output files: `direction.md` and, for `RENDER`, `creative-spec.json` (or
|
|
19
|
+
`slideshow-spec.json` for a slideshow), plus `image-prompts.json` when `media.json` has `generate: true`.
|
|
19
20
|
- Never call Conductor or any MCP tool.
|
|
20
21
|
|
|
21
22
|
## direction.md
|
|
@@ -23,8 +24,11 @@ send to the draft renderer, so it has to be valid on the first try.
|
|
|
23
24
|
Follow the shot-list template: shot list table (time, visual, on-screen text, VO or audio, asset),
|
|
24
25
|
safe zones, sound, and the mapping to the Creative's fields. Keep text out of the platform's UI
|
|
25
26
|
(top ~14%, bottom ~35%, ~6% at each side for 9:16). For a `FILM` idea, include capture settings and
|
|
26
|
-
the ready-to-paste text.
|
|
27
|
-
|
|
27
|
+
the ready-to-paste text. Pick the layout per
|
|
28
|
+
`references/layout-guide.md`: `bleed` by default when 9:16 is enabled, `card` when the photo must stay
|
|
29
|
+
readable or a face would sit under the copy. Never `stacked` or `split` when 9:16 is enabled. Build the
|
|
30
|
+
spec so it would score 3 or more on each criterion in `references/design-critique.md`; the orchestrator
|
|
31
|
+
runs that critique on the render.
|
|
28
32
|
|
|
29
33
|
### Publish assets
|
|
30
34
|
|
|
@@ -35,7 +39,26 @@ licence (original audio, a Commercial Music Library track, a library track with
|
|
|
35
39
|
clip's own sound, or none). Never pick a trending track a business account cannot license. Alt
|
|
36
40
|
text also goes in the spec's `altText`. The orchestrator merges this section into `publish-pack.md`.
|
|
37
41
|
|
|
38
|
-
##
|
|
42
|
+
## slideshow-spec.json (a slideshow RENDER, instead of creative-spec.json)
|
|
43
|
+
|
|
44
|
+
Use this when the idea is an organic slideshow (the orchestrator tells you, and gives you
|
|
45
|
+
`references/slideshow-formats.md`; read it first). Write one JSON object, `{"id", "slides": [...]}`,
|
|
46
|
+
and nothing else; the orchestrator passes it to `compose_slideshow`. Rules:
|
|
47
|
+
|
|
48
|
+
- Follow the post anatomy: slide 1 a human tension, middle slides the value, slide N-1 the app on a
|
|
49
|
+
real screenshot (a `screen` slide, optionally with `callouts`) with one plain mention, the last slide
|
|
50
|
+
a save or follow ask. 5 to 7 slides. Choose each slide's `type` for its job from the table.
|
|
51
|
+
- Photos and the screenshot are paths from `media.json` or local files the person supplied; never
|
|
52
|
+
invent a path. Use a real screenshot only. If `generate: true`, write `image-prompts.json` for the
|
|
53
|
+
photo slots (`{"photo": "...", "beats": {"2": "<prompt for slide 2>"}}`, as below) and leave those
|
|
54
|
+
paths to the orchestrator.
|
|
55
|
+
- Text: at most 8 words a line and 3 lines a box, the first text of each slide has `y`, and every
|
|
56
|
+
text sits between y 16 and 75. No logo, badge or offer on any slide. No invented numbers.
|
|
57
|
+
- In `direction.md`, the Publish assets section still lists the cover frame (card 1), alt text for
|
|
58
|
+
every slide in order, and the sound. Alt text describes each slide for someone who cannot see it.
|
|
59
|
+
- Do not use `creative-spec.json` for it, and do not write layout, theme or headline fields.
|
|
60
|
+
|
|
61
|
+
## creative-spec.json (RENDER only; STILL and MOTION)
|
|
39
62
|
|
|
40
63
|
A single JSON object whose keys are exactly `preview_creative_draft` inputs, and nothing else:
|
|
41
64
|
|
|
@@ -50,7 +73,7 @@ A single JSON object whose keys are exactly `preview_creative_draft` inputs, and
|
|
|
50
73
|
orchestrator fills them. Otherwise, if none is available, stop and say so in `direction.md` rather
|
|
51
74
|
than writing a spec the preview will refuse.
|
|
52
75
|
- `typeOverrides` when text must read bigger or smaller on a placement, and `layoutOverrides` for
|
|
53
|
-
photo height
|
|
76
|
+
photo height, and `focalOverride` for the focal point. Don't change layout just to resize text.
|
|
54
77
|
- `layout` and `theme` only from what the brand kit and the creative registry support (the
|
|
55
78
|
orchestrator lists the options in your instructions). A theme is valid only for layouts that
|
|
56
79
|
support it.
|
|
@@ -36,8 +36,13 @@ Inputs: `brief.md`, `judging/scores.md`, `ideas/pool.md`, the finalist images un
|
|
|
36
36
|
Output: `decision.md`.
|
|
37
37
|
|
|
38
38
|
1. For each rendered finalist, look at its image and judge it as the rubric says (phone-size
|
|
39
|
-
readability, hook legibility in 1 s, safe zones, brand look)
|
|
40
|
-
|
|
39
|
+
readability, hook legibility in 1 s, safe zones, brand look) and score the craft criteria in
|
|
40
|
+
`references/design-critique.md` (1 to 5 each, absolute, one compact line per finalist): the six
|
|
41
|
+
STILL criteria for a still, ad or motion piece, the five slideshow criteria (slide 1 stops the
|
|
42
|
+
scroll, each slide earns the next swipe, people would save or send it, the app shows up as the
|
|
43
|
+
payoff not the opener, native look) for a slideshow. Never score a slideshow on the STILL set, which
|
|
44
|
+
rewards product and logo prominence an organic post should not have. Note
|
|
45
|
+
anything the text round could not see. A generated photo that reads as AI-made (waxy surfaces, duplicated or merged objects,
|
|
41
46
|
warped hands, mushy background; see `references/realism-check.md`) counts against that finalist:
|
|
42
47
|
say so explicitly.
|
|
43
48
|
2. Compare each pair of finalists twice, once in each order, reasoning first. A split is a tie.
|
|
@@ -38,11 +38,17 @@ ideas from your angle, not a safe average.
|
|
|
38
38
|
social-proof).
|
|
39
39
|
5. Tag each idea `RENDER` or `FILM` honestly, using `references/capabilities.md` and `media.json`.
|
|
40
40
|
It is `RENDER` only if every card or frame fits what Conductor can render with media that exists:
|
|
41
|
-
one photo plus text per card,
|
|
42
|
-
|
|
41
|
+
one photo plus text per card, a native slideshow (photos, app screenshots, Notes pages, chats, tier
|
|
42
|
+
lists and similar slides), or motion text over a photo or clip. That covers motion text, which
|
|
43
|
+
needs no filming. There are no type-only stills, voiceover or screen recordings.
|
|
43
44
|
Footage, a person on camera or spoken lines make it `FILM`. For `RENDER`, fill the draft-spec
|
|
44
45
|
block, use only media ids that exist, and respect the copy rules and accent-phrase requirement.
|
|
45
|
-
6.
|
|
46
|
+
6. For an organic post, a `RENDER` idea is usually a slideshow. Name its archetype (saved-but-never-used
|
|
47
|
+
guilt, doubt or conflict, stop doing X, numbered list, tier list, Notes-app plan, POV, before/after;
|
|
48
|
+
see `references/slideshow-formats.md` when given) and the slide count (5 to 7), and put the slide 1
|
|
49
|
+
hook in the on-screen hook. Slide 1 is a human tension, never a feature, the app name or a price.
|
|
50
|
+
Put these in the Concept line; the draft spec may then be "slideshow, <archetype>, <n> slides".
|
|
51
|
+
7. Do not name your angle inside an idea. Do not rank your ideas.
|
|
46
52
|
|
|
47
53
|
## Finish
|
|
48
54
|
|
|
@@ -31,15 +31,23 @@ anything.
|
|
|
31
31
|
against the copy rules and write the result in the final section.
|
|
32
32
|
5. Keep captions within each platform's limits as given in `platform-specs.md`; include only the
|
|
33
33
|
platforms the brief asks for. In the **Publish copy** section also write, per platform: the
|
|
34
|
-
caption with the search keyword in its opening words (and in the on-screen text and speech)
|
|
35
|
-
|
|
34
|
+
caption with the search keyword in its opening words (and in the on-screen text and speech) and a
|
|
35
|
+
call to action that names the app so it can be searched ("search <App> in the App Store"; a
|
|
36
|
+
caption cannot hold a clickable link), 3 to 5 hashtags, a pinned-comment line (where to get it,
|
|
37
|
+
"link in bio"), and a reminder to repost to Instagram Stories with a link sticker. Note the
|
|
38
|
+
branded-content or "your brand" setting the post needs. Then write 2 or 3 alternate hooks,
|
|
36
39
|
each tagged with a hookType (question, bold-claim, pattern-interrupt, relatable-pain, how-to,
|
|
37
40
|
number, curiosity-gap, story, social-proof) that differs from the main hook's type. The
|
|
38
41
|
orchestrator merges this section into `publish-pack.md`. Every factual claim must be one the
|
|
39
42
|
brief, knowledge or approved lines confirm; soften or drop any other.
|
|
40
|
-
6. For a
|
|
43
|
+
6. For a slideshow, the script is slide by slide: slide 1 the human tension, middle slides the value,
|
|
44
|
+
slide N-1 the app on a real screenshot with one plain mention ("I keep mine in <App>"), the last
|
|
45
|
+
slide a save or follow ask. At most 8 words a line and 3 lines a box. No logo, badge or offer on a
|
|
46
|
+
slide; the offer goes in the caption. Invent no numbers: a count must come from the brief or
|
|
47
|
+
knowledge, or be left out.
|
|
48
|
+
7. For a `FILM` idea, write for a person on camera: spoken lines in natural speech, with the
|
|
41
49
|
on-screen text separate. For a `RENDER` idea, keep the text short enough to read at phone size.
|
|
42
|
-
|
|
50
|
+
8. When the orchestrator passes user change requests, apply them and keep everything else.
|
|
43
51
|
|
|
44
52
|
## Finish
|
|
45
53
|
|
|
@@ -11,7 +11,9 @@ One command from "I need a post for X" to a committed Creative (or, for an idea
|
|
|
11
11
|
a shoot-ready script). The team chooses the idea itself; the only point where it stops for you is the
|
|
12
12
|
final draft preview. This command gathers the request and hands off to the `conductor-content-studio`
|
|
13
13
|
skill (`Skill(skill: "conductor-content-studio")`), which runs the stages. Don't re-derive what that
|
|
14
|
-
skill covers.
|
|
14
|
+
skill covers. For an organic post the team's renderable ideas are native slideshows (built with
|
|
15
|
+
`compose_slideshow`, posted as a CAROUSEL Creative); stills and motion pieces stay for ideas that
|
|
16
|
+
are one image, an ad, or a motion piece.
|
|
15
17
|
|
|
16
18
|
## Trigger
|
|
17
19
|
|
|
@@ -6,8 +6,8 @@ allowed-tools: mcp__conductor__*, AskUserQuestion, Skill, WebFetch, Read, Glob,
|
|
|
6
6
|
|
|
7
7
|
# /conductor:creative
|
|
8
8
|
|
|
9
|
-
Guided entry point for making a Post's artwork — a static ad, a short video,
|
|
10
|
-
already on hand. Keeps the interaction short: a couple of questions, then it hands off to the
|
|
9
|
+
Guided entry point for making a Post's artwork — a static ad, a short video, an organic slideshow,
|
|
10
|
+
or posting a video already on hand. Keeps the interaction short: a couple of questions, then it hands off to the
|
|
11
11
|
`conductor-creative` skill (`Skill(skill: "conductor-creative")`) for the actual writing, rendering
|
|
12
12
|
and judgment calls. This command doesn't replace that skill's guidance — read it once invoked and
|
|
13
13
|
follow it.
|
|
@@ -31,7 +31,7 @@ continue with Step 2.
|
|
|
31
31
|
"question": "What would you like to create?",
|
|
32
32
|
"header": "Creative Type",
|
|
33
33
|
"options": [
|
|
34
|
-
{"label": "Static ad", "description": "A photo + headline,
|
|
34
|
+
{"label": "Static ad", "description": "A photo + headline ad with the brand layout, at every enabled placement (not for organic posts)"},
|
|
35
35
|
{"label": "Short video from a photo", "description": "Animated text over a still photo or a moving background"},
|
|
36
36
|
{"label": "Video from a clip I have", "description": "Animate the brand layout using an existing video as background"},
|
|
37
37
|
{"label": "Post an existing video as-is", "description": "No brand layout — the finished clip goes up directly"},
|
|
@@ -51,6 +51,10 @@ Keep this light — one or two follow-up questions, not a form:
|
|
|
51
51
|
already on hand before asking for a new photo; no photo fits and none is offered? Offer to generate one with
|
|
52
52
|
`generate_creative_image` (needs an OpenAI API key, or the person's yes to use Codex). A new photo is a local file path (a URL: download it
|
|
53
53
|
to a local file first) — it is only uploaded after the person approves the preview.
|
|
54
|
+
- **Organic post or slideshow** (the person says so under "Other", or describes a TikTok or Instagram
|
|
55
|
+
photo post that is not an ad): skip the brand layouts. Hand off to the `conductor-creative` skill's
|
|
56
|
+
slideshow path (`references/slideshow-formats.md`): `compose_slideshow`, then a CAROUSEL preview,
|
|
57
|
+
approval and commit. Ask for the topic, the photos and a real app screenshot, if there is one.
|
|
54
58
|
- **Video from a clip I have / Post an existing video as-is**: call
|
|
55
59
|
`list_creative_media({kind: "VIDEO"})` first; if nothing fits, ask for a local path (a URL:
|
|
56
60
|
download it first). A clip used as a MOTION background is previewed from the local file too.
|
|
@@ -74,11 +78,13 @@ Nothing is saved to Conductor until the person approves what they have seen:
|
|
|
74
78
|
2. Read `passed` and `checks` in the result first. The preview runs the full render's per-placement
|
|
75
79
|
checks (text spill, platform safe zones such as TikTok/Reels' overlay band, contrast, fonts,
|
|
76
80
|
images), so a design that would fail the real render shows up here. If `passed` is false, fix the
|
|
77
|
-
design (shorter copy, `layoutOverrides`, `typeOverrides`, a
|
|
81
|
+
design (shorter copy, `layoutOverrides`, `typeOverrides`, a better layout per the skill's
|
|
82
|
+
`references/layout-guide.md`) and preview again
|
|
78
83
|
**before** showing it or asking anything — never offer Approve & upload while a blocking check
|
|
79
84
|
fails, and tell the person what you changed. (Bigger or smaller text on one placement is
|
|
80
85
|
`typeOverrides`, e.g. `{"9x16": [96]}`, not a layout switch.)
|
|
81
|
-
3. Once `passed` is true,
|
|
86
|
+
3. Once `passed` is true, score the image with the skill's `references/design-critique.md` (fix anything
|
|
87
|
+
below 3, at most twice), then show it with the compact score line and ask with
|
|
82
88
|
`AskUserQuestion`: Approve & upload / Request changes / Stop. On changes, preview again; on
|
|
83
89
|
Stop, leave it — the draft stays local under `.conductor/drafts/`.
|
|
84
90
|
4. On approval, `commit_creative_draft` with the `draftDir`. It uploads the files, saves the
|
|
@@ -92,8 +98,8 @@ Nothing is saved to Conductor until the person approves what they have seen:
|
|
|
92
98
|
|
|
93
99
|
When the Creative came from a content-team run, or the person points to a publish pack, its
|
|
94
100
|
`publish-pack.json` carries per-destination captions and publish options (disclosure, AI label,
|
|
95
|
-
cover frame, alt text) and a
|
|
101
|
+
cover frame, alt text) and a pinned comment. Look for `.conductor/content-runs/*/publish-pack.json`
|
|
96
102
|
whose `creativeId` matches the Creative. When you go on to create or edit the Post with the
|
|
97
103
|
`conductor-publisher` skill, pre-fill each target's `captionOverride` and options from it, show them
|
|
98
|
-
in the confirmation table and let the person change them. Mention the
|
|
99
|
-
|
|
104
|
+
in the confirmation table and let the person change them. Mention the pinned comment (and, for a post that promotes an app, the Stories repost with a link sticker) as manual
|
|
105
|
+
steps, since Conductor cannot post comments. The Post still goes through the normal review.
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
---
|
|
2
2
|
name: conductor-content-studio
|
|
3
|
-
description: Runs a small team of subagents (strategist, three ideators, creative director, scriptwriter, art director) that brainstorm competing short-form social post ideas, pick one with a bias-resistant judging method, write the script and visual direction, run pre-publish checks (disclosure, AI label, music, claims, third-party material, accessibility), and render a local draft with a publish pack the publisher later pre-fills from. The person approves at the preview; only then is the Creative uploaded to Conductor. Ideas are either RENDER (a still, carousel, story sequence or motion text that Conductor can render) or FILM (needs real filming, so the output is a shoot-ready script). A run ends at a committed Creative or a shoot script and never creates a Post. Use when asked to come up with, choose and make a short-form post (TikTok, Reels, Shorts, Facebook) from a goal rather than from a given headline, or via /conductor:content. For a headline or photo already chosen, use conductor-creative instead.
|
|
3
|
+
description: Runs a small team of subagents (strategist, three ideators, creative director, scriptwriter, art director) that brainstorm competing short-form social post ideas, pick one with a bias-resistant judging method, write the script and visual direction, run pre-publish checks (disclosure, AI label, music, claims, third-party material, accessibility), and render a local draft with a publish pack the publisher later pre-fills from. The person approves at the preview; only then is the Creative uploaded to Conductor. Ideas are either RENDER (a native slideshow for organic posts, or a still, carousel, story sequence or motion text that Conductor can render) or FILM (needs real filming, so the output is a shoot-ready script). A run ends at a committed Creative or a shoot script and never creates a Post. Use when asked to come up with, choose and make a short-form post (TikTok, Reels, Shorts, Facebook) from a goal rather than from a given headline, or via /conductor:content. For a headline or photo already chosen, use conductor-creative instead.
|
|
4
4
|
user-invocable: true
|
|
5
5
|
allowed-tools: mcp__conductor__*, Agent, AskUserQuestion, Skill, Read, Write, Glob, Grep, WebFetch
|
|
6
6
|
---
|
|
@@ -30,6 +30,10 @@ Each exists because of a known failure mode (sources are in `references/rubric.m
|
|
|
30
30
|
- **One gate.** The person is asked only at the final draft preview. Do not stop to ask which idea.
|
|
31
31
|
The one exception is S5b's disclosure question, asked only when disclosure cannot be inferred.
|
|
32
32
|
- **A run never creates a Post.** It ends at a committed Creative (or a shoot script).
|
|
33
|
+
- **Organic posts default to a slideshow.** A `RENDER` idea for an organic post (not an ad) is a
|
|
34
|
+
native-looking slideshow built by `compose_slideshow` and posted as a CAROUSEL; see
|
|
35
|
+
`references/slideshow-formats.md`. STILL and MOTION stay for ideas that really are one still, an
|
|
36
|
+
ad, or a motion piece. A slideshow never carries a logo, badge or offer.
|
|
33
37
|
|
|
34
38
|
Cost: a run is about 8 subagent calls (strategist 1, ideators 3, director 2, scriptwriter 1, art
|
|
35
39
|
director 1), and only the strategist and the director use the top model tier. `--quick` runs two
|
|
@@ -44,6 +48,7 @@ folder; give subagents absolute paths.
|
|
|
44
48
|
|
|
45
49
|
```
|
|
46
50
|
request.md run.json brief.md script.md direction.md creative-spec.json decision.md
|
|
51
|
+
slideshow-spec.json (S5, instead of creative-spec.json, when the winner is a slideshow)
|
|
47
52
|
image-prompts.json (S5, only when generation is enabled)
|
|
48
53
|
context/ brand-kit.json knowledge.md performance.md learning.md media.json
|
|
49
54
|
ideas/ angle-1.md angle-2.md angle-3.md pool.md winner.md
|
|
@@ -122,7 +127,8 @@ Write `request.md` (the request verbatim, `--quick` if given, the platform if st
|
|
|
122
127
|
not, using views, engagement rate, average view percentage, 72 h views and any 3-second or
|
|
123
128
|
completion figure the insights expose, and any experiment result. Name what you could not
|
|
124
129
|
measure. With no earlier run or no data, write exactly "No data yet".
|
|
125
|
-
6. Every renderable idea needs a photo (or, for MOTION, a clip
|
|
130
|
+
6. Every renderable idea needs a photo (or, for MOTION, a clip; for a slideshow's payoff slide, a real
|
|
131
|
+
app screenshot, so ask for one if the media has none); see `references/capabilities.md`. If
|
|
126
132
|
`media.json` has no unblocked image and no clip, ask once with `AskUserQuestion` before ideating:
|
|
127
133
|
**Give a photo or two** (local paths; record them in `context/media.json` as `local` entries the
|
|
128
134
|
art director may use via `photoPath`), **Generate a photo** (record `"generate": true` in
|
|
@@ -175,7 +181,10 @@ before each score, and names the top 3. Store the finalists in `run.json`.
|
|
|
175
181
|
|
|
176
182
|
### S3b — Finalist drafts (you; skipped in `--quick`)
|
|
177
183
|
|
|
178
|
-
|
|
184
|
+
A slideshow finalist has no cheap draft: judge it from its idea text (archetype and slide-by-slide
|
|
185
|
+
text) unless its slides can be composed from photos already on hand. In that case write its spec to
|
|
186
|
+
`<run>/drafts/finalist-<id>/slideshow-spec.json`, call `compose_slideshow` with `draftDir` set to
|
|
187
|
+
that folder and judge from the first slide and the `checks`. For every other `RENDER` finalist, call `preview_creative_draft` with that idea's draft spec (map the
|
|
179
188
|
draft-spec block to the tool's fields) and `draftDir` set to `<run>/drafts/finalist-<id>`. Use
|
|
180
189
|
`full: false`. Record each draft dir in `run.json`. For a finalist whose draft spec carries an
|
|
181
190
|
`imagePrompt`, call `generate_creative_image({prompt})` first, run the realism check
|
|
@@ -195,7 +204,7 @@ path as `photoPath` (generate only for finalists, not the whole pool).
|
|
|
195
204
|
|
|
196
205
|
Same agent, mode `visual-round`. Inputs: `brief.md`, `judging/scores.md`, `ideas/pool.md`, the
|
|
197
206
|
finalist image paths (`drafts/finalist-<id>/sheet.jpg`; the director views them with Read), the
|
|
198
|
-
rubric and `references/compliance-checklist.md`. It compares finalists pairwise in both orders, runs
|
|
207
|
+
rubric, `references/design-critique.md` and `references/compliance-checklist.md`. It compares finalists pairwise in both orders, runs
|
|
199
208
|
the devil's-advocate pass against the leader (including the veto questions: an unsubstantiated
|
|
200
209
|
claim, trademark or parody risk, a representation or sensitivity problem, disclosure needed), and
|
|
201
210
|
writes `decision.md`: the winner, why it won, the runners-up, one line per cut idea and the veto
|
|
@@ -212,13 +221,22 @@ which idea to take; show the decision at the gate.
|
|
|
212
221
|
First copy the winning idea's block from `ideas/pool.md` to `ideas/winner.md`. Then
|
|
213
222
|
`Agent(subagent_type: "conductor-content-scriptwriter")` with `brief.md`, `decision.md`, `ideas/winner.md`, `references/script-template.md`, `references/platform-specs.md` and
|
|
214
223
|
`references/publish-pack-template.md`. Besides the script it writes the Publish copy section: a
|
|
215
|
-
caption per platform
|
|
224
|
+
caption per platform whose call to action names the app so it can be searched, hashtags, a pinned
|
|
225
|
+
comment and 2 to 3 alternate hooks with a hookType each.
|
|
216
226
|
For a `MOTION` winner tell it that the render shows one headline and one optional body, animated
|
|
217
227
|
by preset, then an end card that holds the final frame: no timed multi-beat on-screen text. Extra
|
|
218
228
|
beats belong in a story or carousel sequence, or in the caption. When merging into `publish-pack.json`, append each platform's hashtags to its `captionOverride`
|
|
219
229
|
(caption, blank line, hashtags): that field is posted verbatim.
|
|
220
230
|
|
|
221
|
-
### S5 — Art director (`RENDER` only) -> `direction.md` + `creative-spec.json`
|
|
231
|
+
### S5 — Art director (`RENDER` only) -> `direction.md` + `creative-spec.json` or `slideshow-spec.json`
|
|
232
|
+
|
|
233
|
+
For a slideshow winner the art director writes `slideshow-spec.json` instead of `creative-spec.json`
|
|
234
|
+
(and no `image-prompts.json` unless it needs a photo made). Give it `references/slideshow-formats.md`,
|
|
235
|
+
`context/media.json`, `script.md`, `references/platform-specs.md`, `references/design-critique.md` and
|
|
236
|
+
`references/publish-pack-template.md`. Photos and the app screenshot are paths from `media.json`
|
|
237
|
+
(download a library image to a local file first) or the person's local files; generated photos go
|
|
238
|
+
through `generate_creative_image` and the realism check first, as below. The rest of this stage is
|
|
239
|
+
the STILL/MOTION path.
|
|
222
240
|
|
|
223
241
|
`Agent(subagent_type: "conductor-content-art-director")` with `script.md`, `context/media.json`,
|
|
224
242
|
`context/brand-kit.json`, `references/shotlist-template.md`, `references/platform-specs.md` and
|
|
@@ -227,14 +245,16 @@ beats belong in a story or carousel sequence, or in the caption. When merging in
|
|
|
227
245
|
licence per platform.
|
|
228
246
|
Tell it the supported options: the kit's enabled placements, and the layouts and themes
|
|
229
247
|
the creative registry supports (at the time of writing `stacked`, `bleed`, `split` dark only, `card` dark or
|
|
230
|
-
light; a refusal from the validator names the valid ones).
|
|
248
|
+
light; a refusal from the validator names the valid ones). Also give it `references/layout-guide.md`
|
|
249
|
+
and `references/design-critique.md`: it picks the layout per the guide, and never `stacked` or `split`
|
|
250
|
+
when 9:16 is enabled. The spec's keys must be exactly
|
|
231
251
|
`preview_creative_draft` inputs. Give it `references/capabilities.md` too. The spec's `name` is the
|
|
232
252
|
idea's title only: never include the internal `I<n>` ID, which means nothing outside this run. When
|
|
233
253
|
`media.json` has `generate: true`, the art director may also write `image-prompts.json`
|
|
234
254
|
(`{"photo": "...", "beats": {"2": "..."}}`) next to the spec. Before S6, call
|
|
235
255
|
`generate_creative_image` for each prompt, run the realism check (`references/realism-check.md`,
|
|
236
256
|
regenerating on a fail) on each image, and set `photoPath` (or
|
|
237
|
-
`sequence[i].photoPath`) in `creative-spec.json
|
|
257
|
+
`sequence[i].photoPath`) in `creative-spec.json` (for a slideshow, the slide's `bg` or `photos[].src` in `slideshow-spec.json`; `beats` keys are slide numbers). A `FILM` winner skips S5 and goes to the shoot
|
|
238
258
|
script, below.
|
|
239
259
|
|
|
240
260
|
### S5b — Publish pack and pre-publish checks (you; RENDER after S5, FILM after the shoot script)
|
|
@@ -243,7 +263,7 @@ Before the gate, so the person sees the settings with the draft. Skip nothing he
|
|
|
243
263
|
does not apply is answered "not applicable".
|
|
244
264
|
|
|
245
265
|
1. Run `references/compliance-checklist.md` against `brief.md`, `script.md`, `direction.md` (or the
|
|
246
|
-
shoot script), `creative-spec.json` and the provenance in `context/media.json` (`aiGenerated`,
|
|
266
|
+
shoot script), `creative-spec.json` (or `slideshow-spec.json`) and the provenance in `context/media.json` (`aiGenerated`,
|
|
247
267
|
`licence`).
|
|
248
268
|
2. Merge the scriptwriter's Publish copy and the art director's Publish assets, add the compliance
|
|
249
269
|
results, and write `publish-pack.md` and `publish-pack.json` in the shapes of
|
|
@@ -267,10 +287,18 @@ does not apply is answered "not applicable".
|
|
|
267
287
|
`draftDir` (the default location is cleaned up automatically after the commit). Record the
|
|
268
288
|
returned `draftDir` as `drafts.final`. Local files (`photoPath`, `clipPath`, `audioPath`) go in
|
|
269
289
|
as given in the spec.
|
|
290
|
+
**Slideshow winner:** `compose_slideshow({specPath})` on `slideshow-spec.json`; while `passed` is
|
|
291
|
+
false, fix the spec and compose again. Then `preview_creative_draft({kind: "CAROUSEL", slidePaths,
|
|
292
|
+
caption, slideAltText})` with the returned `slidePaths` (`9x16` for TikTok, `4x5` for Instagram),
|
|
293
|
+
the caption from `publish-pack.json` and the alt text from `direction.md`. Record the preview's
|
|
294
|
+
`draftDir` as `drafts.final`.
|
|
270
295
|
2. Eye-check the image. Load the `conductor-creative` skill with `Skill(skill: "conductor-creative")` and
|
|
271
|
-
apply its review guidance (step 7 and the silent-failures list)
|
|
296
|
+
apply its review guidance (step 7 and the silent-failures list), scoring the draft with
|
|
297
|
+
`references/design-critique.md` (the slideshow criteria for a slideshow, the STILL criteria
|
|
298
|
+
otherwise; any criterion below 3 is fixed before anyone sees it). For a STILL: does the headline read at a
|
|
272
299
|
glance, does the accent phrase fall on the words that carry the meaning, is copy on a face, are
|
|
273
|
-
the safe zones clear.
|
|
300
|
+
the safe zones clear. For a slideshow: does slide 1 stop the scroll, is the text inside y 16 to 75,
|
|
301
|
+
is the screenshot real, and is there no logo, badge or offer. Also read the result's `checks`/`passed`: while any blocking check fails,
|
|
274
302
|
fix it (`layoutOverrides` for photo height, `typeOverrides` for text size, a shorter body) and
|
|
275
303
|
re-preview **before** showing anything, and never offer Approve & upload while one still fails.
|
|
276
304
|
Revise at most twice beyond that; if it is still wrong, show it honestly and say what is off.
|
|
@@ -278,12 +306,13 @@ does not apply is answered "not applicable".
|
|
|
278
306
|
`drafts/finalist-<id>/sheet.jpg`; say when a runner-up could not be rendered; with `--quick` no
|
|
279
307
|
sheets exist, so show runners-up by title and tag only), and a short decision
|
|
280
308
|
summary (the winner and why, the runners-up, what was cut and why), and a short pack summary: the
|
|
281
|
-
disclosure setting (which toggle, or none), the AI-label setting, the sound and its licence,
|
|
282
|
-
the alternate hooks with their types. For each generated photo, add its one-line realism verdict.
|
|
309
|
+
disclosure setting (which toggle, or none), the AI-label setting, the sound and its licence, the
|
|
310
|
+
pinned comment and the Stories link-sticker repost, and the alternate hooks with their types. For each generated photo, add its one-line realism verdict.
|
|
283
311
|
4. Ask with `AskUserQuestion`, header "Draft", with exactly these four options. `AskUserQuestion`
|
|
284
312
|
allows at most four, so never add one per runner-up, and **Stop** must always be among them:
|
|
285
313
|
- **Approve & upload**: the question says approving also confirms the photo carries no burned-in
|
|
286
|
-
text, logo or claim
|
|
314
|
+
text, logo or claim (for a slideshow: every number and claim on the slides is real and the
|
|
315
|
+
screenshot is genuine). Ask for provenance (`source`, `licence`, and whether it is AI-generated)
|
|
287
316
|
only for local files the preview lists in `provenance.needed` (a file from
|
|
288
317
|
`generate_creative_image` already carries it), then
|
|
289
318
|
`commit_creative_draft({draftDir, photoChecked: true, source, licence, aiGenerated})` (provenance
|
|
@@ -309,7 +338,7 @@ does not apply is answered "not applicable".
|
|
|
309
338
|
and `references/shotlist-template.md`, output `shoot/shoot-script.md`: script, shot list, capture
|
|
310
339
|
settings (9:16, 30 fps, sound on), B-roll, safe zones, and the on-screen text and captions ready to
|
|
311
340
|
paste. It also writes the Publish assets section there. It writes no `creative-spec.json` here.
|
|
312
|
-
2. Run S5b. The shoot script gets the publish pack too (captions, hashtags,
|
|
341
|
+
2. Run S5b. The shoot script gets the publish pack too (captions, hashtags, pinned comment,
|
|
313
342
|
disclosure, music and its licence) so it can be applied when the clip is posted later; append a
|
|
314
343
|
short "Publish pack" section to `shoot/shoot-script.md`, and leave `creativeId` null in
|
|
315
344
|
`publish-pack.json`.
|
|
@@ -324,9 +353,9 @@ Keep it short:
|
|
|
324
353
|
|
|
325
354
|
- the decision summary: winner, one line on why, the runners-up;
|
|
326
355
|
- the Creative display id (and the readiness gaps, if any) or the shoot doc path;
|
|
327
|
-
- `publish-pack.md` (captions, hashtags,
|
|
356
|
+
- `publish-pack.md` (captions, hashtags, pinned comment, disclosure and AI-label settings, sound,
|
|
328
357
|
alternate hooks); say the publisher will pre-fill the Post from `publish-pack.json`, and that the
|
|
329
|
-
|
|
358
|
+
pinned comment, the link in bio and the Stories repost with a link sticker are manual steps;
|
|
330
359
|
- next steps:
|
|
331
360
|
- make the Post with `/conductor:creative` or the `conductor-publisher` skill (the Creative is
|
|
332
361
|
not on a Post yet, because a run never creates one);
|
|
@@ -347,4 +376,5 @@ Keep it short:
|
|
|
347
376
|
## Related
|
|
348
377
|
|
|
349
378
|
- `conductor-creative`: the single-Creative skill whose checklist S6 uses.
|
|
379
|
+
- `references/slideshow-formats.md`: the slideshow spec, archetypes and text rules.
|
|
350
380
|
- `conductor-publisher`: takes the committed Creative onto a Post, pre-filled from `publish-pack.json`.
|
|
@@ -27,6 +27,12 @@ needs changing until it fits.
|
|
|
27
27
|
(`slides`, with `default` as the fallback) and each must match its placement's aspect. Reels,
|
|
28
28
|
Stories and YouTube cannot take it.
|
|
29
29
|
|
|
30
|
+
- **Slideshow** (organic posts): native-looking slides made locally by `compose_slideshow` (photo with
|
|
31
|
+
text, app screenshot with callouts, Notes page, chat, photo grid, stacked before/after, tier list,
|
|
32
|
+
week planner, receipt, placard), then posted as a CAROUSEL (see `references/slideshow-formats.md`).
|
|
33
|
+
This is the default for an organic `RENDER` post. It needs real photos and, for the payoff slide,
|
|
34
|
+
a real app screenshot; it takes no Brand Kit layout. Ads keep using STILL.
|
|
35
|
+
|
|
30
36
|
## Hard requirements (a draft is refused without them)
|
|
31
37
|
|
|
32
38
|
- **A photo.** Every STILL needs one, and so does a carousel or story, whose beats inherit the main
|
|
@@ -44,6 +50,8 @@ needs changing until it fits.
|
|
|
44
50
|
|
|
45
51
|
- **Layouts:** `stacked`, `bleed` and `split` are dark only; `card` is dark or light. The photo and
|
|
46
52
|
text arrangement is fixed per layout.
|
|
53
|
+
- Pick the layout with `references/layout-guide.md`: `bleed` by default when 9:16 is enabled, `card`
|
|
54
|
+
when the photo must stay readable. `stacked` and `split` are for 4:5 only.
|
|
47
55
|
- **`stacked` drops the body at 4:5.** That includes every card of a 4:5 carousel, and it happens
|
|
48
56
|
silently. If a body line carries meaning (a claim, the product name, the accent phrase), use
|
|
49
57
|
`card`, `split` or `bleed`, or move the line into the headline.
|
|
@@ -63,13 +71,13 @@ When `media.json` has `generate: true`, one generated photo per card or beat (or
|
|
|
63
71
|
background) is renderable. It comes from `generate_creative_image` (OpenAI Images API, or Codex with the person's yes):
|
|
64
72
|
photographic, no text, logos or screens in the picture, and no depiction of the product unless a real
|
|
65
73
|
reference image is supplied. It is `aiGenerated: true`. Run `references/realism-check.md` on every one. Collages, charts and text-in-image stay out.
|
|
66
|
-
Ask for a mid-tone scene: under `bleed`, a dark or night scene can render almost black. For a dark scene, use `card
|
|
74
|
+
Ask for a mid-tone scene: under `bleed`, a dark or night scene can render almost black. For a dark scene, use `card`.
|
|
67
75
|
|
|
68
76
|
## What it cannot do (so the idea needs filming or a rethink)
|
|
69
77
|
|
|
70
78
|
- Footage of people, places, cooking or the product in use. Screen recordings of an app.
|
|
71
79
|
- Voiceover or spoken dialogue: MOTION has music or clip sound, not speech.
|
|
72
|
-
-
|
|
73
|
-
|
|
80
|
+
- Charts, illustrations, or text inside a generated image. In a STILL each card is one photo plus
|
|
81
|
+
text; grids, stacks and collages exist only as slideshow slides (above).
|
|
74
82
|
- Anything that relies on a product claim the brief can't confirm. That's a copy problem, not a
|
|
75
83
|
rendering one: soften the claim or drop the idea.
|
|
@@ -0,0 +1,52 @@
|
|
|
1
|
+
<!-- Keep identical to the copy in the other creative skill (conductor-creative / conductor-content-studio references/design-critique.md). Skills install independently and cannot reference each other. -->
|
|
2
|
+
# Design critique
|
|
3
|
+
|
|
4
|
+
Passing checks does not make an ad good. `checks` and `passed` cover text spill, safe zones, contrast
|
|
5
|
+
and fonts. This critique covers craft, and it runs before the person sees a draft.
|
|
6
|
+
|
|
7
|
+
Look at the preview image at phone size. Vision models miss small typography details, so don't
|
|
8
|
+
claim to have judged kerning or hairline spacing; judge what is visible at a glance.
|
|
9
|
+
|
|
10
|
+
## Which criteria apply
|
|
11
|
+
|
|
12
|
+
Stills, ads and brand-layout creatives use the six criteria below. An organic slideshow (a
|
|
13
|
+
`compose_slideshow` CAROUSEL) uses the five slideshow criteria further down instead, because it is
|
|
14
|
+
meant to look native, not like an ad.
|
|
15
|
+
|
|
16
|
+
## Score each criterion 1 to 5 (stills and ads)
|
|
17
|
+
|
|
18
|
+
Score each one on its own, against the description, not against another draft.
|
|
19
|
+
|
|
20
|
+
1. **Thumb-stop**: would it stop a scroll? One clear focal point and a hook you get in a second.
|
|
21
|
+
2. **Headline size and position**: readable at thumbnail size, not crammed into a corner or a thin strip.
|
|
22
|
+
3. **Empty space**: no dead slab or empty panel. Space that is there on purpose is fine; a leftover band is not.
|
|
23
|
+
4. **Photo and copy together**: copy is not on the subject's face, and the photo carries the concept rather than decorating it.
|
|
24
|
+
5. **Product shown**: the product, or the result it gives, is visible. A stock pose or an empty scene scores 1 or 2.
|
|
25
|
+
6. **Not a banner ad**: it does not read as logo + badge + tagline stacked on a picture.
|
|
26
|
+
|
|
27
|
+
## Act on the scores
|
|
28
|
+
|
|
29
|
+
- Any criterion below 3: fix it (copy, layout, `focalOverride`, `typeOverrides`, the photo) and
|
|
30
|
+
render again. At most two fix rounds.
|
|
31
|
+
- Still below 3 after that: show the best version with one honest line on what is weak.
|
|
32
|
+
- Comparing variants: score each on its own first. Use a side-by-side comparison only to break a
|
|
33
|
+
tie, and judge it in both orders (first vs. second, then second vs. first); if the order changes
|
|
34
|
+
the winner, call it a tie. Pairwise judging favours whichever is shown first.
|
|
35
|
+
|
|
36
|
+
## Show the scores
|
|
37
|
+
|
|
38
|
+
One compact line next to the preview, for example:
|
|
39
|
+
`Craft: thumb-stop 4, headline 4, space 3, photo+copy 4, product 3, not-banner 4. Weakest: product is implied, not shown.`
|
|
40
|
+
|
|
41
|
+
## Slideshow criteria (organic slideshows)
|
|
42
|
+
|
|
43
|
+
Look at the contact sheet: slide 1 at full size, then the slides in order. Score each 1 to 5.
|
|
44
|
+
|
|
45
|
+
1. **Slide 1 stops the scroll**: a human tension or a result people want, in a line a stranger gets in a second. A feature, the app name or a price scores 1 or 2.
|
|
46
|
+
2. **Each slide earns the next swipe**: every slide has one job and a reason to go on; no filler slide, no slide that repeats the last.
|
|
47
|
+
3. **People would save or send it**: the middle slides hold something to keep or to send to a named person.
|
|
48
|
+
4. **The app is the payoff, not the opener**: it appears on slide N-1 on a real screenshot with one plain mention, never on slide 1 and never as a logo, badge or offer.
|
|
49
|
+
5. **Native look**: text boxes and photos feel like a person's post. Banner-ad signs (a logo lockup, a badge, a tagline stack, a call-to-action button) score 1 or 2.
|
|
50
|
+
|
|
51
|
+
Act on the scores as below. Show them in one line, for example:
|
|
52
|
+
`Slideshow: scroll-stop 4, next-swipe 3, save/send 4, app-as-payoff 4, native 5. Weakest: slide 3 repeats slide 2.`
|
|
@@ -13,7 +13,7 @@ Heading: `## Idea <n>: <title>` where the title is 8 words or fewer.
|
|
|
13
13
|
- **Hook, spoken:** <max 12 words, or "none (text only)">
|
|
14
14
|
- **Hook, on screen:** <max 12 words>
|
|
15
15
|
- **Hook type:** <question | bold-claim | pattern-interrupt | relatable-pain | how-to | number | curiosity-gap | story | social-proof>
|
|
16
|
-
- **Format:** <still | carousel | story sequence | motion text over photo | motion text over clip | talking head | demo | skit | other>
|
|
16
|
+
- **Format:** <slideshow | still | carousel | story sequence | motion text over photo | motion text over clip | talking head | demo | skit | other>
|
|
17
17
|
- **Platform:** <primary platform, then any others>
|
|
18
18
|
- **Tag:** RENDER | FILM
|
|
19
19
|
- **Why this audience cares:** <max 30 words, tied to the brief's tension>
|
|
@@ -21,7 +21,7 @@ Heading: `## Idea <n>: <title>` where the title is 8 words or fewer.
|
|
|
21
21
|
|
|
22
22
|
Draft spec (RENDER only):
|
|
23
23
|
|
|
24
|
-
- kind: STILL | MOTION
|
|
24
|
+
- kind: STILL | MOTION | slideshow (organic post: name the archetype and slide count, 5 to 7, instead of the fields below)
|
|
25
25
|
- layout: <a layout the brand kit and registry support>
|
|
26
26
|
- theme: dark | light
|
|
27
27
|
- headline: <with exactly one *accent phrase* if the kit requires it>
|
|
@@ -33,7 +33,7 @@ Draft spec (RENDER only):
|
|
|
33
33
|
- motion: <MOTION only: preset, durationSec, background>
|
|
34
34
|
```
|
|
35
35
|
|
|
36
|
-
The idea's format maps to the publish pack's `tags.format` as: still -> `still`; carousel ->
|
|
36
|
+
The idea's format maps to the publish pack's `tags.format` as: slideshow -> `carousel`; still -> `still`; carousel ->
|
|
37
37
|
`carousel`; story sequence -> `story`; motion text over photo or clip -> `motion`; talking head,
|
|
38
38
|
demo, skit or other -> `film`.
|
|
39
39
|
|
|
@@ -0,0 +1,52 @@
|
|
|
1
|
+
<!-- Keep identical to the copy in the other creative skill (conductor-creative / conductor-content-studio references/layout-guide.md). Skills install independently and cannot reference each other. -->
|
|
2
|
+
# Layout guide
|
|
3
|
+
|
|
4
|
+
Pick the layout on purpose. A Creative has one layout across all its placements (9:16, 4:5, 1:1), so
|
|
5
|
+
choose it for the hardest placement, usually 9:16. Never rely on the registry default: it is
|
|
6
|
+
`stacked`, which is the wrong pick whenever 9:16 is enabled.
|
|
7
|
+
|
|
8
|
+
These notes come from rendering the same photo and headline in every layout with
|
|
9
|
+
`preview_creative_draft`.
|
|
10
|
+
|
|
11
|
+
## Default
|
|
12
|
+
|
|
13
|
+
- 9:16 enabled: **`bleed`**. Use **`card`** when the photo itself must stay bright and readable.
|
|
14
|
+
- Only 4:5 and 1:1: `split` or `card` both work.
|
|
15
|
+
|
|
16
|
+
## The layouts
|
|
17
|
+
|
|
18
|
+
- **`bleed`**: full-frame photo with a large headline. It is the only layout that fills 9:16 with a
|
|
19
|
+
big headline. Dark theme only (`theme: "light"` is refused). Its scrim darkens most of the photo on
|
|
20
|
+
9:16 and 4:5, so choose a photo with a bright, high-contrast subject. On 4:5 the headline sits at
|
|
21
|
+
the top, so a face in the top third of the photo lands under it; on 1:1 the headline is small at the
|
|
22
|
+
bottom.
|
|
23
|
+
- **`card`**: the photo as a framed card with a large headline beneath. On 9:16 the bottom third is
|
|
24
|
+
an empty dark panel. Use it when the photo must stay readable, such as a product screen that
|
|
25
|
+
bleed's scrim would hide.
|
|
26
|
+
- **`split`**: good on 4:5, with a big headline under the photo. Bad on 9:16: the photo sits in the
|
|
27
|
+
top third, a large dark panel is left empty, the headline is small, and it failed the 9:16 bottom
|
|
28
|
+
safe-zone check. Don't use it when 9:16 is enabled.
|
|
29
|
+
- **`stacked`**: on 9:16 the photo band leaves an empty dark slab and the headline is forced small
|
|
30
|
+
(the 430 px bottom safe zone is solid fill). Don't use it when 9:16 is enabled. Fine for 4:5.
|
|
31
|
+
|
|
32
|
+
## 1:1 headline size
|
|
33
|
+
|
|
34
|
+
On 1:1 every layout auto-fits the headline small (about 60 px at most). Pin it with `typeOverrides`,
|
|
35
|
+
e.g. `{"1x1": [88]}`; 92 passed the checks with an eight-word headline. A pinned size is not shrunk to
|
|
36
|
+
fit, so read `checks` again after pinning.
|
|
37
|
+
|
|
38
|
+
## Logo
|
|
39
|
+
|
|
40
|
+
`lockup: "chip"` puts the logo on a white pill. Use it when the photo behind the logo is busy.
|
|
41
|
+
|
|
42
|
+
## Photo and layout together
|
|
43
|
+
|
|
44
|
+
`focalOverride` (e.g. `{"4x5": "50% 80%"}`) only moves the subject when the crop leaves spare room:
|
|
45
|
+
a photo much taller than the placement's frame, or `card`'s wide strip on 1:1. A photo already close
|
|
46
|
+
to the placement's shape (a 4:5-ish photo on 4:5) has nothing to shift, and the override does nothing.
|
|
47
|
+
Check the next preview rather than assuming it worked.
|
|
48
|
+
|
|
49
|
+
If the subject's face would sit under the headline: try `focalOverride` where there is room, else pick
|
|
50
|
+
a photo with the subject lower in the frame, else use `card` (and pin `card`'s 1:1 crop with
|
|
51
|
+
`focalOverride` so the face isn't cut off). If two fix rounds still leave a weak frame, the photo is
|
|
52
|
+
usually the problem: say so and offer a different one.
|
|
@@ -28,9 +28,14 @@ Run: <runId> Creative: <display id, or "not yet committed"> Tags: hook <type
|
|
|
28
28
|
|
|
29
29
|
## <Platform> (<feed | reel | story>, as the destination's formats allow)
|
|
30
30
|
- **Caption:** <final text; the search keyword is in the first line, within the first ~125
|
|
31
|
-
characters;
|
|
31
|
+
characters; the call to action names the app so it is searchable ("search <App> in the App
|
|
32
|
+
Store"), because captions hold no clickable link; within the platform's limit>
|
|
32
33
|
- **Hashtags:** <3 to 5>
|
|
33
|
-
- **
|
|
34
|
+
- **Pinned comment (post manually):** <text, such as where to get the app, or "none">. Conductor cannot post comments.
|
|
35
|
+
- **Link in bio / Stories:** <"link in bio is set" and, for a post that promotes an app or offer,
|
|
36
|
+
"repost to Instagram Stories with a link sticker">. Manual steps.
|
|
37
|
+
- **Branded-content setting:** <the platform's branded-content or "your brand" toggle for a brand's
|
|
38
|
+
own promotional post, and the AI label when a generated photo is used>
|
|
34
39
|
- **Cover frame:** <video: timestamp in seconds and what is on it; carousel: card 1; still: n/a>
|
|
35
40
|
- **Alt text:** <one line per image or card, in order>
|
|
36
41
|
- **On-screen text placement:** <where each line sits, inside the safe zone>
|
|
@@ -94,7 +99,7 @@ Rules:
|
|
|
94
99
|
`list_publish_targets`. `captionOverride` is ignored on a story.
|
|
95
100
|
- `captionOverride` is the **complete text that gets posted**: the caption, then a blank line, then that
|
|
96
101
|
platform's hashtags. The publisher posts it as-is, so hashtags listed only in `publish-pack.md` never
|
|
97
|
-
reach the platform. The
|
|
102
|
+
reach the platform. The pinned comment is the one thing that stays out: it is posted by hand.
|
|
98
103
|
- `options` holds real option keys only, in the camelCase the destination reports under
|
|
99
104
|
`optionKeys`. Never add a key that is not in this table, and never `privacyLevel` (the person
|
|
100
105
|
picks it from the account's allowed levels at publish time):
|
|
@@ -58,6 +58,12 @@ Look at the actual image file, not the text description. Check:
|
|
|
58
58
|
- Hook legibility in 1 second: is the first thing seen the hook, in high contrast?
|
|
59
59
|
- Safe zones: text out of the platform UI (see `platform-specs.md`), nothing sitting on a face.
|
|
60
60
|
- Brand look: tokens, font, logo lockup, accent phrase on the words that carry the meaning.
|
|
61
|
+
- Craft: score the criteria in `design-critique.md`. For a still, ad or brand-layout creative, the
|
|
62
|
+
six STILL criteria (thumb-stop, headline size and position, empty space, photo and copy together,
|
|
63
|
+
product shown, not a banner ad). For a slideshow, the five slideshow criteria (slide 1 stops the
|
|
64
|
+
scroll, each slide earns the next swipe, people would save or send it, the app shows up as the
|
|
65
|
+
payoff not the opener, native look, not a banner ad), each 1 to 5. A finalist with a criterion below 3
|
|
66
|
+
that the layout could have fixed (`stacked` or `split` on 9:16, a dead panel) counts against it.
|
|
61
67
|
- Realism: a generated photo that looks AI-made (waxy, duplicated objects, warped hands) counts against the finalist.
|
|
62
68
|
- Whether the picture and the words do different jobs (photo shows, words claim) rather than
|
|
63
69
|
repeating each other.
|