ppt-template-adaptation

Name: ppt-template-adaptation
Author: CxyZyr

15stars1forks

How to install ppt-template-adaptation?

cxyzyr/pptx-template-skills/ppt-template-adaptation

$npx -y skills add CxyZyr/PPTX-Template-Skills --skill ppt-template-adaptation

Installs into the current project.

Summary

Use to generate or debug a new deck from an existing PPTX template and a parsed spec.json. Enforces the second half of the parse→generate pipeline: select pages, build one global outline, author content_plan.json, fill text while preserving parsed template style, replace icons/lo

Files · 1

View on GitHub

SKILL.md

1	# PPT Template Adaptation
2
3	Use this skill when the task is to rewrite an existing PPT template into a new deck while preserving the
4	template's visual logic. It consumes `spec.json` from `pptx-template-parsing`; do not start from a failed
5	generated deck or manually inferred shape indices.
6
7	This skill is intentionally self-contained. Do not read old external handbooks from previous projects.
8
9	## What this skill enforces
10
11	- Parse first, generate second. If a slot, card count, or style is missing, fix parsing before adding
12	generation fallback logic.
13	- Work page by page. Incremental runs keep original slide indices stable until the final full build.
14	- Keep one global outline as the source of truth for contents pages, section dividers, and section titles.
15	- Fill text before icons, logo, and images.
16	- Preserve parsed template style: geometry, alignment, theme colors, paragraph/run structure, and number
17	badges unless the user asks for redesign.
18	- Replace icons and images only when semantics require it and a valid slot exists.
19	- Validate each page before moving on.
20
21	## Inputs
22
23	- Original template `.pptx`.
24	- Parsed `spec.json` from `pptx-template-parsing`.
25	- A `content_plan.json` with `spec`, `output`, `asset_dir`, optional `brand`, `keep_slides`, optional
26	`outline`, and `pages[]`.
27
28	## Required workflow
29
30	1. If no current `spec.json` exists, run `pptx-template-parsing` first.
31	2. Read the parser `summary.txt`; choose template slides whose roles and slot counts match the needed
32	story. Avoid chart/table slides unless chart/table data will be supplied.
33	3. Scaffold a plan:
34
35	```bash
36	python3 skills/ppt-template-adaptation/scripts/scaffold_content_plan.py \
37	--spec workspace/specs/<template>/spec.json \
38	--out workspace/plans/<deck>.content_plan.json \
39	--output-pptx workspace/out/<deck>.pptx
40	```
41
42	4. Edit `keep_slides`, `outline`, and `pages[]`. Map each page to existing slots only: titles to
43	`title`, free labels to `labels[]`, body boxes to `body[]`, repeated structures to `cards[]`,
44	semantic glyphs to `cards[].icon`, and photos/screenshots to `images[]`.
45	5. Validate global text coherence:
46
47	```bash
48	python3 skills/ppt-template-adaptation/scripts/validate_content_plan.py \
49	--plan workspace/plans/<deck>.content_plan.json
50	```
51
52	6. Fill one or a few original slide indices at a time:
53
54	```bash
55	python3 skills/ppt-template-adaptation/scripts/apply_content_plan.py \
56	--plan workspace/plans/<deck>.content_plan.json \
57	--pages 0,1 \
58	--render
59	```
60
61	7. Review the rendered page when visual review is available. If the user will do visual review, provide
62	the output files and continue with structural checks; do not use image tools when the user forbids it.
63	8. Repeat the page loop until selected pages pass. Then run a full build without `--pages` so unused
64	slides are dropped according to `keep_slides`.
65
66	## Debugging order
67
68	1. Inspect parser output first: `summary.txt`, `spec.json` `fill_plan`, card count, `style.paragraphs`,
69	font availability, and visual obstacles.
70	2. If parsing is wrong, fix `pptx-template-parsing`, rerun parsing, and rebuild the plan or affected
71	pages.
72	3. If parsing is right but generation breaks style or layout, fix `ppt-template-adaptation` application
73	logic.
74	4. Use `--no-clear-unfilled`, `--skip-global-text-check`, or `--skip-image-resolve` only for deliberate
75	diagnostics. Do not use them to declare a page complete.
76
77	## Reference read order
78
79	Read only the references needed for the current task:
80
81	- `references/workflow.md` — complete page-by-page workflow and acceptance criteria.
82	- `references/text-coherence.md` — contents/section outline consistency and text fitting.
83	- `references/icon-fill.md` — semantic icon replacement rules.
84	- `references/image-fill.md` — conservative image replacement and Tavily usage.
85
86	## Example scripts
87
88	- `scripts/scaffold_content_plan.py`
89	- `scripts/validate_content_plan.py`
90	- `scripts/apply_content_plan.py`
91	- `scripts/render_pages.py`
92	- `scripts/tavily_image_service.py`
93
94	## Usage rule
95
96	Do not jump to the next page until the current page has passed validation. For visual validation, either
97	review the render when allowed or hand the render artifact to the user for manual review.

People who liked this also viewed

6 picks

Type

ppt-template-adaptation

15stars1forks

View on GitHub

How to install ppt-template-adaptation?

cxyzyr/pptx-template-skills/ppt-template-adaptation

$npx -y skills add CxyZyr/PPTX-Template-Skills --skill ppt-template-adaptation

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# PPT Template Adaptation
2
3	Use this skill when the task is to rewrite an existing PPT template into a new deck while preserving the
4	template's visual logic. It consumes `spec.json` from `pptx-template-parsing`; do not start from a failed
5	generated deck or manually inferred shape indices.
6
7	This skill is intentionally self-contained. Do not read old external handbooks from previous projects.
8
9	## What this skill enforces
10
11	- Parse first, generate second. If a slot, card count, or style is missing, fix parsing before adding
12	generation fallback logic.
13	- Work page by page. Incremental runs keep original slide indices stable until the final full build.
14	- Keep one global outline as the source of truth for contents pages, section dividers, and section titles.
15	- Fill text before icons, logo, and images.
16	- Preserve parsed template style: geometry, alignment, theme colors, paragraph/run structure, and number
17	badges unless the user asks for redesign.
18	- Replace icons and images only when semantics require it and a valid slot exists.
19	- Validate each page before moving on.
20
21	## Inputs
22
23	- Original template `.pptx`.
24	- Parsed `spec.json` from `pptx-template-parsing`.
25	- A `content_plan.json` with `spec`, `output`, `asset_dir`, optional `brand`, `keep_slides`, optional
26	`outline`, and `pages[]`.
27
28	## Required workflow
29
30	1. If no current `spec.json` exists, run `pptx-template-parsing` first.
31	2. Read the parser `summary.txt`; choose template slides whose roles and slot counts match the needed
32	story. Avoid chart/table slides unless chart/table data will be supplied.
33	3. Scaffold a plan:
34
35	```bash
36	python3 skills/ppt-template-adaptation/scripts/scaffold_content_plan.py \
37	--spec workspace/specs/<template>/spec.json \
38	--out workspace/plans/<deck>.content_plan.json \
39	--output-pptx workspace/out/<deck>.pptx
40	```
41
42	4. Edit `keep_slides`, `outline`, and `pages[]`. Map each page to existing slots only: titles to
43	`title`, free labels to `labels[]`, body boxes to `body[]`, repeated structures to `cards[]`,
44	semantic glyphs to `cards[].icon`, and photos/screenshots to `images[]`.
45	5. Validate global text coherence:
46
47	```bash
48	python3 skills/ppt-template-adaptation/scripts/validate_content_plan.py \
49	--plan workspace/plans/<deck>.content_plan.json
50	```
51
52	6. Fill one or a few original slide indices at a time:
53
54	```bash
55	python3 skills/ppt-template-adaptation/scripts/apply_content_plan.py \
56	--plan workspace/plans/<deck>.content_plan.json \
57	--pages 0,1 \
58	--render
59	```
60
61	7. Review the rendered page when visual review is available. If the user will do visual review, provide
62	the output files and continue with structural checks; do not use image tools when the user forbids it.
63	8. Repeat the page loop until selected pages pass. Then run a full build without `--pages` so unused
64	slides are dropped according to `keep_slides`.
65
66	## Debugging order
67
68	1. Inspect parser output first: `summary.txt`, `spec.json` `fill_plan`, card count, `style.paragraphs`,
69	font availability, and visual obstacles.
70	2. If parsing is wrong, fix `pptx-template-parsing`, rerun parsing, and rebuild the plan or affected
71	pages.
72	3. If parsing is right but generation breaks style or layout, fix `ppt-template-adaptation` application
73	logic.
74	4. Use `--no-clear-unfilled`, `--skip-global-text-check`, or `--skip-image-resolve` only for deliberate
75	diagnostics. Do not use them to declare a page complete.
76
77	## Reference read order
78
79	Read only the references needed for the current task:
80
81	- `references/workflow.md` — complete page-by-page workflow and acceptance criteria.
82	- `references/text-coherence.md` — contents/section outline consistency and text fitting.
83	- `references/icon-fill.md` — semantic icon replacement rules.
84	- `references/image-fill.md` — conservative image replacement and Tavily usage.
85
86	## Example scripts
87
88	- `scripts/scaffold_content_plan.py`
89	- `scripts/validate_content_plan.py`
90	- `scripts/apply_content_plan.py`
91	- `scripts/render_pages.py`
92	- `scripts/tavily_image_service.py`
93
94	## Usage rule
95
96	Do not jump to the next page until the current page has passed validation. For visual validation, either
97	review the render when allowed or hand the render artifact to the user for manual review.

People who liked this also viewed

6 picks

Type