AI turns documents or topics into real, native PowerPoint decks—with native shapes, transitions and animations, data-backed charts and tables on demand, audio narration from speaker notes, and support for your own .pptx templates. · by Hugo He
docs(executor): calibrate text width once per role instead of measuring every line
hugohe3acba8696ac44
diff --git a/skills/ppt-master/references/executor-base.md b/skills/ppt-master/references/executor-base.mdindex 03e93a14..965f3eed 100644--- a/skills/ppt-master/references/executor-base.md+++ b/skills/ppt-master/references/executor-base.md@@ -159,7 +159,7 @@ Before drawing each page, look up its entry in `page_rhythm` (key format `P<NN>`- **Element grouping (Mandatory)**: wrap each logical Slide-local body unit in a descriptive, page-unique top-level `<g id>`. Every visible direct root `<g>` except a compact helper-authored preset atom declares root-coordinate `data-pptx-bounds="x y width height"`; that text-free atom stays top-level when standalone, uses `data-pptx-frame`, and never carries bounds. Frame/native coordinates do not replace bounds on any other group, and placeholder bounds also supply the slot frame. Nested groups need no bounds and any such values are ignored. Checker fails ordinary root-group overlap exceeding `1px` on both axes; structured slots, structural-role groups, and off-canvas Morph staging groups are exempt, but structured Slide-local groups are not. Checker compares root bounds with the `viewBox`, recursively checks estimable text—including both shared multiline forms—against its root module with DrawingML wrapping headroom, and independently checks every estimable visible text carrier against the page without that headroom: through `1px` is ignored; module overflow warns through `5%` and fails above it, while larger page overflow always fails. Unestimable visible text receives an advisory warning. Only a wholly off-canvas direct-root Morph endpoint may set `data-pptx-morph-staging="true"`; keep its text inside its own module bounds, use an explicit pair when Morph remains enabled, and never use the marker for partial overflow. Images, shapes, paths, `<use>`, effects, and object frames remain geometrically free. Flat pages use ordinary groups; structured slots already qualify, while titles, direct Master/Layout atoms, and canvas-level static framing may remain root primitives. On flat pages, give a root background image or full-canvas scrim/decoration rectangle a stable `id` plus `data-pptx-role="background"` / `"decoration"`; never wrap it only to silence the advisory.- **Reference — not a constraint**: top-level groups set semantic and automatic-animation granularity, but they may contain descriptive nested `<g>` edit groups when the page has meaningful internal subunits. Nested groups need no bounds and create no automatic animation step; use or omit them from the page's actual editing semantics, with no default pattern, depth, or quota.-- **Default — size `data-pptx-bounds` as the intended module zone, not a glyph box (may skip when no text is estimable)**: make the zone as generous as the canvas and sibling layout allow, without overlapping another module zone. An untransformed line spans `y - 0.85 × font_size` to `y + 0.35 × font_size`; width uses the shared SVG-to-PPTX per-run estimate and safety headroom. The same estimator is callable before writing: `python3 ${SKILL_DIR}/scripts/text_measure.py measure|wrap|box ...` measures lines, wraps one paragraph to a max width as ready `<tspan>` rows, and computes a text block's module zone. If text does not fit, first expand a zone that has unused non-overlapping space; otherwise reflow or adapt. Larger bounds do not repair off-canvas text.+- **Default — size `data-pptx-bounds` as the intended module zone, not a glyph box (may skip when no text is estimable)**: make the zone as generous as the canvas and sibling layout allow, without overlapping another module zone. An untransformed line spans `y - 0.85 × font_size` to `y + 0.35 × font_size`; width uses the shared SVG-to-PPTX per-run estimate and safety headroom. The same estimator is callable before writing: `python3 ${SKILL_DIR}/scripts/text_measure.py measure|wrap|box ...` measures lines, wraps one paragraph to a max width as ready `<tspan>` rows, and computes a text block's module zone; one calibration per role serves the deck, and later calls are for lines that approach a limit, batched through `--stdin`. If text does not fit, first expand a zone that has unused non-overlapping space; otherwise reflow or adapt. Larger bounds do not repair off-canvas text.- **Spec adherence**: follow binding color, canvas, typography, identity, resource, and template anchors; apply layout and other Reference directions under §2.1 without turning them into locks- **Template structure**: inherit the native visual framework only for `template_reuse_scope: mirror|layout`; `style` uses the flat route- **Main-agent ownership**: SVG generation must run in the main agent (not sub-agents) — pages share upstream context for cross-page visual continuitydiff --git a/skills/ppt-master/scripts/prompt_audit_manifest.json b/skills/ppt-master/scripts/prompt_audit_manifest.jsonindex 54acc6f5..54c49cfa 100644--- a/skills/ppt-master/scripts/prompt_audit_manifest.json+++ b/skills/ppt-master/scripts/prompt_audit_manifest.json@@ -1047,7 +1047,7 @@"route.generate.planning-ai"],"files": [],- "max_tokens": 120000+ "max_tokens": 125000},"stage.generate.executor.chart": {"description": "Conditional value-driven chart execution rules.",diff --git a/skills/ppt-master/workflows/generate-pptx.md b/skills/ppt-master/workflows/generate-pptx.mdindex 7298a20b..4784483f 100644--- a/skills/ppt-master/workflows/generate-pptx.md+++ b/skills/ppt-master/workflows/generate-pptx.md@@ -726,7 +726,7 @@ Run the command unfiltered (no `tail`/`head`/`grep`). Review the complete P01 is| Signal | Reading ||---|---|-| Two or more issues share a category and direction | Method-level bias — resolve it to the authoritative rule before P02; a correction fitted to the observed offset only patches this sample. For text extents that rule is the shared estimator, exposed as `python3 ${SKILL_DIR}/scripts/text_measure.py measure|wrap|box` — measure or wrap lines before writing them instead of estimating by hand |+| Two or more issues share a category and direction | Method-level bias — resolve it to the authoritative rule before P02; a correction fitted to the observed offset only patches this sample. For text extents that rule is the shared estimator, exposed as `python3 ${SKILL_DIR}/scripts/text_measure.py measure|wrap|box` — calibrate each role once; measure only lines near a limit (`--stdin` batches) || One isolated issue tied to this page's structure | Page-local — fix and continue || A recurring element appears for the first time (page furniture, caption format, section numbering, accent discipline) | It will be copied to every later page — confirm its semantics now |diff --git a/skills/ppt-master/workflows/profiles/quick-generate.md b/skills/ppt-master/workflows/profiles/quick-generate.mdindex ed025dd1..8c7c2ba7 100644--- a/skills/ppt-master/workflows/profiles/quick-generate.md+++ b/skills/ppt-master/workflows/profiles/quick-generate.md@@ -690,7 +690,7 @@ representative line per recurring text role with`python3 ${SKILL_DIR}/scripts/text_measure.py measure` and keep "≈ N chars perW px" in active context; the final checker's module check adds DrawingMLwrapping headroom, so characters × font size underestimates by roughly 15%.-Wrap long paragraphs with `text_measure.py wrap` instead of counting by hand.+Those per-role constants cover most lines; measure only lines that approach a limit, many at once through `--stdin`, and `wrap` only a genuinely long paragraph.**Generation pacing**: the current main agent hand-writes the SVG roster inorder. Use P01 to calibrate visual identity and cover-specific expression; use