Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add calesthio/generative-media-skills --skill captions-media-accessibilitygit clone --depth 1 https://github.com/calesthio/generative-media-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/calesthio/generative-media-skills/captions-media-accessibility)<a href="https://agentmods.dev/skills/calesthio/generative-media-skills/captions-media-accessibility"><img src="https://agentmods.dev/badge/skills/calesthio/generative-media-skills/captions-media-accessibility/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/calesthio/generative-media-skills/captions-media-accessibility"><img src="https://agentmods.dev/badge/skills/calesthio/generative-media-skills/captions-media-accessibility.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high Anti-Refusal · line 292 Skill instructs the agent to omit warnings, disclaimers, or ethical commentary. Stripping safety caveats hides risk from the user and is a common jailbreak preamble.Fix: Remove instructions that suppress warnings, disclaimers, or ethical commentary. Let the agent surface safety-relevant caveats to the user.
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00094 | $0.06002 |
| Opus 5 | $0.00047 | $0.03001 |
| Sonnet 5 | $0.00019 | $0.01200 |
| Haiku 4.5 | $0.00009 | $0.00600 |
Grade A, and why
captions-media-accessibility scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 399 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Captions and media accessibility direction
Treat captions and accessibility tracks as production assets, not polish. Build them from the script, audio mix, edit, localization plan, and platform delivery target. If the user asks for legal compliance, say what standards you are using and recommend review by qualified accessibility/legal counsel; do not promise legal compliance from a generated file alone.
Start by classifying the media
Before writing or exporting captions, identify:
- Content type: prerecorded video with audio, live/near-live video, audio-only, video-only, or silent social cut.
- Audience need: captions for deaf/hard-of-hearing viewers, translated subtitles for hearing viewers, SDH, transcript, descriptive transcript, audio description, sign language, or platform-default auto captions.
- Delivery target: social burn-in, web player sidecar, broadcast/OTT, LMS/training portal, internal review, localized package, or archival master.
- Accessibility risk: essential visual text, charts, screen recordings, multiple speakers, overlapping dialogue, music/lyrics, sound-driven story beats, flashing/strobe, fast kinetic typography, low contrast, small mobile screens, or platform UI overlays.
- Final-source truth: the locked audio mix and final picture. Generated scripts, TTS drafts, and ASR transcripts are starting points, not the authority after edits.
Documented facts to preserve
Use these as factual constraints when relevant, citing the source in your output or handoff notes when the claim matters.
- WCAG 2.2 requires captions for prerecorded audio in synchronized media at Level A (SC 1.2.2), captions for live synchronized audio at Level AA (SC 1.2.4), and audio description for prerecorded video at Level AA (SC 1.2.5). WCAG Level A allows either audio description or a media alternative for prerecorded synchronized video under SC 1.2.3. Source: W3C WCAG 2.2, verified 2026-07-10.
- WCAG 2.2 flashing guidance says content must not flash more than three times in any one-second period unless below the general/red flash thresholds. Source: W3C Understanding SC 2.3.1, verified 2026-07-10.
- WCAG contrast minimum for text is 4.5:1, with 3:1 for large-scale text. This is a web-text criterion; burned-in captions are video pixels, but use it as the minimum design target for caption readability unless a stricter platform/spec applies. Source: W3C WCAG 2.2 SC 1.4.3, verified 2026-07-10.
- Section508.gov distinguishes closed captions, open captions, subtitles, and transcripts: closed captions can be turned on/off; open captions are permanent in the video; subtitles usually translate dialogue for hearing audiences and do not normally include speaker IDs or non-speech audio; transcripts are not time-coded. Source: Section508.gov synchronized media guidance, verified 2026-07-10.
- U.S. FCC caption quality rules for covered television programming define quality around accuracy, synchronicity, completeness, and placement; captions should include relevant nonverbal information, be legible, and avoid blocking essential visual content. Source: 47 CFR Sec. 79.1 via Legal Information Institute/eCFR text, verified 2026-07-10. This skill does not determine whether a specific project is covered by FCC rules.
- YouTube Help listed SRT and SBV/SubViewer as basic formats and WebVTT/TTML/SAMI/RealText as advanced formats; basic SRT/SBV files must be plain UTF-8 and YouTube notes that styling support is limited. Source: YouTube Help, verified 2026-07-10.
- Vimeo Help stated that Vimeo supports SRT and WebVTT for captions/subtitles, recommends WebVTT, and requires UTF-8 encoding for special characters. Source: Vimeo Help, verified 2026-07-10.
- WebVTT is a UTF-8 timed-text format used with media tracks; it begins with
WEBVTT, contains timed cues, and can be used for captions, subtitles, descriptions, chapters, or metadata depending on player support. Sources: W3C/MDN and Library of Congress WebVTT documentation, verified 2026-07-10. - W3C WebVTT syntax requires UTF-8, a
WEBVTTfile signature, cue timings with period milliseconds, ordered cue start times, and end times greater than start times. Cue overlap is allowed by the format for captions/subtitles, but many caption delivery workflows still reject or flag overlaps. Source: W3C WebVTT Candidate Recommendation Draft, verified 2026-07-11. - SRT is a widely used but lightly standardized plain-text sidecar convention made of repeated blocks: numeric counter, start/end timing line, caption text, and a blank-line separator. Common delivery practice uses
HH:MM:SS,mmm --> HH:MM:SS,mmmtiming and UTF-8 for modern platform upload, even though legacy SRT files may use other encodings. Source: Library of Congress SRT format description plus platform guidance already cited here, verified 2026-07-11. - Netflix timed-text guidance is platform-specific, not universal law. Its public partner help guidance includes 2-line maximums, language-specific style guides, minimum/maximum event durations in general requirements, and U.S. English SDH conventions for bracketed speaker IDs/sound effects. Source: Netflix Partner Help Center, verified 2026-07-10.
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 399 lines · 94 tokens per session scan A 9196bf1b4547
captions-media-accessibility is a skill published in the GitHub repository calesthio/generative-media-skills (170 stars, last pushed 1mo ago), licensed MIT. It adds 94 tokens to every session and 6,002 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
cliptalk-cover-director
Produces evidence-backed cover candidates and reviewable cover variants for a ClipTalk video. Use when the user asks for a cover, poster frame, thumbnail, or multiple cover directions; do not use for timeline editing or social-video reframing.
cliptalk-smart-reframe
Creates a subject-aware, time-varying crop track and a review-only social-format preview from an accepted ClipTalk cut. Use for automatic vertical, square, or portrait reframing; do not use for a fixed manual crop or before content editing is accepted.
cliptalk-audio-polish-mixer
Produces a task-local audio-polished preview or output from an existing cut, applying loudness normalization, optional noise reduction, fades, muting, and voice-first mix policy.
cliptalk-broll-overlay-editor
Adds relevant B-roll or cutaway overlays to an existing timeline while preserving the primary audio and making the result reviewable before final export.
cliptalk-caption-layout-director
Lays out subtitles for an existing ClipTalk timeline, including top/bottom placement, safe margins, readable wrapping, bilingual order, and review rendering.
cliptalk-content-extractor
Locates and assembles source passages matching a semantic request. Use for extracting explanations, topics, quotes, demonstrations, or other specifically described content.