@lincyaw/skills 1.0.0-rc.82 → 1.0.0-rc.84

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/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@lincyaw/skills",
3
- "version": "1.0.0-rc.82",
3
+ "version": "1.0.0-rc.84",
4
4
  "description": "Presmium's bundled authoring, animation, presentation and plugin-development skills for coding agents",
5
5
  "license": "Apache-2.0",
6
6
  "repository": {
@@ -36,7 +36,7 @@
36
36
  ],
37
37
  "dependencies": {
38
38
  "jiti": "^2.7.0",
39
- "@lincyaw/kernel": "1.0.0-rc.82"
39
+ "@lincyaw/kernel": "1.0.0-rc.84"
40
40
  },
41
41
  "presmium": {
42
42
  "skills": [
package/skills/index.json CHANGED
@@ -12,8 +12,8 @@
12
12
  "references/publishing.md",
13
13
  "references/writing.md"
14
14
  ],
15
- "version": "1.0.0-rc.82",
16
- "hash": "a3a7551e1f169ecdc26572a7981a6f69a21f10e3bdd441da879088b731a36011"
15
+ "version": "1.0.0-rc.84",
16
+ "hash": "9046070b8e51de7c039773774b62766bcb2843733d445f645d66d5598a49c24f"
17
17
  },
18
18
  {
19
19
  "name": "presmium-animate",
@@ -545,8 +545,8 @@
545
545
  "references/team.md",
546
546
  "references/timeline.md"
547
547
  ],
548
- "version": "1.0.0-rc.82",
549
- "hash": "129ae2099f0a5313ae0c3f09e7408fb81701e70a05f413e465aa911ec1548cad"
548
+ "version": "1.0.0-rc.84",
549
+ "hash": "81c195a4d2fffb626f1c6c318293808476e4f1219b07d91c3a3fba46bf171649"
550
550
  },
551
551
  {
552
552
  "name": "presmium-develop",
@@ -556,7 +556,7 @@
556
556
  "references/plugin.md",
557
557
  "references/testing.md"
558
558
  ],
559
- "version": "1.0.0-rc.82",
559
+ "version": "1.0.0-rc.84",
560
560
  "hash": "96422f5cef6a97bba23c9c67b7fb0ee491b18f01f789e262ca24b274e7fabaa1"
561
561
  }
562
562
  ]
@@ -6,7 +6,9 @@ Put narration text, pronunciation substitutions and cue nodes in `narrations`. C
6
6
 
7
7
  `production.defaults`, section overrides and step overrides decide whether narration starts with or after the animation, minimum holds, trailing pauses and voice source. Cue delay belongs to `production.cueTimings`. Recording or media clips reference an asset, source range and step attachment. Runtime playback start times and durations are derived.
8
8
 
9
- **Scene states follow the voice.** A scene definition's `durations` give each state's length on the reading page, on slides stepped by hand, in `capture_slides` frames, and as the step's minimum hold in the plan. While the narrated script plays (the presentation's lecture, the Video Production preview, the video export), state k lasts until the voice of the first step that shows it ends, never less than `durations[k]`; the pause and hold after the voice rest on the state's final frame. A step whose narration starts after the animation (`after-animation`), a step without voice and a recorded lecture keep the authored durations. The scene receives the durations it plays with as `durations`, `at(k)` and `total`.
9
+ **Scene states follow the voice automatically.** A scene definition's `durations` give each state's fallback length on the reading page, on slides stepped by hand and in `capture_slides` frames. During the narrated script (the presentation's lecture, the Video Production preview and the video export), a scene whose TTS starts with its animation fills the actual voice length of the first step that shows it, whether shorter or longer than the fallback. Explicit `production.minimumHoldMs` still holds the final frame, and `trailingPauseMs` still adds its pause; do not duplicate preview `durations` into minimum holds. A step whose narration starts after the animation (`after-animation`), a step without voice, a recording and a page with a continuous next-step cue keep authored animation timing. Other fixed-duration components retain their own minimum.
10
+
11
+ The scene receives its own clock as `durations`, `at(k)` and `total`, not the physical voice milliseconds: each state's build budget is at least its authored duration and extends for a longer voice. The viewer maps physical time into that state's scene clock independently. A shorter voice therefore speeds the whole state up, including fixed-length tweens, without cutting its rest frame or shifting the next state's boundary. Use these scene-clock values consistently; do not independently measure wall time or retime them after synthesis.
10
12
 
11
13
  So place every beat inside a state at a fraction of that state, `at(k) + f * durations[k]` (GSAP seconds: divide by 1000), with f where its key word falls in the sentence; keep fixed-length moves (an entrance, a draw) anchored to such a beat. Never place a beat at a fixed time or retype a state's length: the beat then lands with its words whatever the voice's length, and `durations` need no rewriting after a voiceover. Set `durations` to the pace the state should have without a voice. `capture_slides` with `at` draws the authored timeline; check the voiced timing in the exported video's frames.
12
14
 
@@ -102,6 +102,8 @@ Completed audio belongs to the exact content and voice configuration that produc
102
102
 
103
103
  Cue content supplies stable targets and actions; `production.cueTimings` owns delays. Timing defaults/overrides decide narration start, minimum hold and trailing pause. Recording/media clips use asset references, explicit source ranges and step attachments. Derived absolute playback times stay in the playback plan.
104
104
 
105
+ Simultaneous TTS automatically paces a `PmScene` to its actual voice length, shrinking or extending the scene without changing its fallback `durations`. Use `minimumHoldMs` only for a deliberate minimum shot length, never to copy a silent preview's duration; an explicit minimum still holds the final picture after the voice. Silence, animation-first narration, recordings, continuous next-step cues and fixed-duration components retain authored timing. See `presmium-animate/references/timeline.md` for the complete scene-clock rules.
106
+
105
107
  Local audio `production.mediaClips` can reference an existing canonical audio asset in the same document. Set `sourceRange` in milliseconds, an attached stable step and offset, `volume` from 0 to 1, and optional `muted`; muting retains the volume. The inspected audio range must fit entirely inside that confirmed step. Adjust trimming, offset or the authored hold when it does not fit; the plan does not retime cues or cross steps automatically. These clips mix with cached speech in preview and video export, without generating or changing captions. Files must be non-hidden, regular local audio; symlinks, remote/shared paths and video are rejected. Use the source-bound `production-asset` reader to inspect a referenced AssetID and the plan diagnostics to confirm availability. Saving JSON does not confirm the actual audio range.
106
108
 
107
109
  To use original audio as narration, use Video’s **Original audio → Use as step voiceover**, or explicitly call the shared `adopt_recording` action with `stepId` and `assetId`; the host pins the saved source and actual SHA-256 internally before approval; see `DocumentRecordingAdoption`. This sets `voiceSource.recording`, never a fabricated media clip. Each effective step plays its selected source range once; without a range it plays the whole file. In Video, **Narration range** edits the already adopted same-asset step with real in/out bounds. Restore full changes pending input and needs explicit Apply. Editing one step does not trim the asset or another step. Without an explicitly applied valid manual timing map, current Narration display text forms one whole-clip caption without word alignment; empty narration produces no caption. Trim keeps the whole authored script as an unaligned step caption and requires review; original ASR word times do not become trimmed alignment. Continuous cross-step recording, internal pauses, non-start cues and conflicting narration starts remain unsupported. Adoption preserves all narration and pronunciation, and does not start ASR, TTS or playback. The accepted version is prepared paused; late replies preserve newer input and microphone captures. Browser-controlled recording evidence does not verify a physical microphone, and MCP recording-client control uses the same mounted candidate through `editor_action`.
@@ -135,6 +137,6 @@ Set the longest the video may be in the lecture renderer options: `metadata.rend
135
137
 
136
138
  A step whose visible blocks are all limited to other forms (`only` without `lecture`, such as a sources page whose text has `only: ["web"]`) and that has no narration, recording or audio clip, and no `{cue}next` of an earlier narration moves into, is not in the lecture: the video, the plan, the captions and the chapters leave it out, and a page whose steps are all like that is left out whole. It needs no production override to hide it.
137
139
 
138
- MCP `get_voiceover_status` returns eight rows by default (up to twenty) within a 7168-character JSON budget. Follow `nextCursor` as `cursor` for further step targets, unbound pages and job summaries; source or preview changes require a fresh first page. Use `get_voiceover_job` for complete task details. Continuation carries its own version binding; no extra verification arguments are needed.
140
+ MCP `get_voiceover_status` automatically prepares the current saved document when its compiled result is missing or stale; do not open a preview or read an IR resource as a separate warm-up. It never generates audio. Source changes during preparation and stale generation revisions remain conflicts. Capabilities and voice catalogues do not compile scenes; job discovery and cancellation remain available without compilation. The status returns eight rows by default (up to twenty) within a 7168-character JSON budget. Follow `nextCursor` as `cursor` for further step targets, unbound pages and job summaries; source or preview changes require a fresh first page. Use `get_voiceover_job` for complete task details. Continuation carries its own version binding; no extra verification arguments are needed.
139
141
 
140
142
  MCP `get_voice_catalog` and `get_voiceover_jobs` also use bounded pages (default eight, maximum twenty). Follow `nextCursor` with the same original filters. Job admission, polling and cancellation summarize large step selections and mark shortened detail. To read exact task fields, use `get_voiceover_job` with `path`, for example `["job","scope","targets"]`; `path: []` selects everything. Small selections return `encoding: "value"`; large ones return JSON-text fragments. Follow the cursor with the same job query, assemble the decoded text, then parse it. The cursor carries the field, offset and consistency check; no extra verification arguments.