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 amazon-pollygit 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/amazon-polly)<a href="https://agentmods.dev/skills/calesthio/generative-media-skills/amazon-polly"><img src="https://agentmods.dev/badge/skills/calesthio/generative-media-skills/amazon-polly/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/amazon-polly"><img src="https://agentmods.dev/badge/skills/calesthio/generative-media-skills/amazon-polly.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
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 →
- medium analysis-evasion · line 1 Suspicious Unicode normalization or mixed-script contentFix: Review the flagged content for security risks. Ensure no credentials, secrets, or sensitive data are exposed.
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 | $0.00092 | $0.05887 |
| Opus 5 | $0.00046 | $0.02943 |
| Sonnet 5 | $0.00018 | $0.01177 |
| Haiku 4.5 | $0.00009 | $0.00589 |
Grade A, and why
amazon-polly 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 9d 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 — 314 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Amazon Polly production TTS
Use Amazon Polly when the job needs AWS-hosted text-to-speech with predictable API operations, IAM governance, lexicons, SSML, and reusable generated audio. Treat Polly as a production TTS service, not a voice-cloning system: it provides AWS-managed voices and engines, not custom voice training or arbitrary speaker imitation.
Facts below were verified from official AWS documentation on 2026-07-10. Re-check AWS docs before relying on volatile facts such as prices, quotas, regions, voice lists, or engine feature compatibility.
First decision: which operation and engine
Choose the operation first, because it defines latency, output custody, and timing options:
SynthesizeSpeech: real-time request/response. Use for short narration lines, UI prompts, voice preview, batch segment synthesis, and speech-mark JSON. Input limit is 6,000 total characters, with no more than 3,000 billed characters; output audio is cut off after 10 minutes. Documented fact: https://docs.aws.amazon.com/polly/latest/dg/limits.htmlStartSpeechSynthesisTask: asynchronous synthesis to Amazon S3, optionally with SNS notification. Use for long articles, audiobook chapters, training modules, and server workflows that should not hold an HTTP audio stream open. Input limit is 200,000 total characters, with no more than 100,000 billed characters. Documented fact: https://docs.aws.amazon.com/polly/latest/APIReference/API_StartSpeechSynthesisTask.htmlStartSpeechSynthesisStream: bidirectional streaming input/output over HTTP/2. Use only when the application needs streamed generative audio while text is still arriving. Documented fact: AWS currently documents this operation as supporting only thegenerativeengine and not JSON speech marks. Source: https://docs.aws.amazon.com/polly/latest/APIReference/API_StartSpeechSynthesisStream.html
Then select the engine:
standard: broadest compatibility, lowest listed commercial price, strong for utility prompts, IVR, high-volume notification audio, and workflows where cost/coverage matter more than expressiveness. AWS describes Standard voices as concatenative synthesis.neural: better naturalness than standard, with all speech marks, many SSML tags, asynchronous synthesis, and selected Newscaster speaking style support. Good default for produced narration when speech marks or SSML control matter.long-form: designed for longer narrative content such as articles, training, and marketing videos. AWS documents only six long-form voices as of 2026-07-10: en-US Danielle, Gregory, Ruth, Patrick and es-ES Alba, Raúl; region availability is US East (N. Virginia) only. Use when long-duration delivery quality matters and the voice/language constraints fit.generative: most conversational and adaptive Polly family, with streaming support and higher humanlike variability. Use for conversational assistant lines, virtual trainers, and ad-like reads when its feature constraints fit. Important documented boundaries: speech marks are not currently available; Newscaster style is not supported; AWS warns the model can vary after updates and includes a hallucination/emergency-stop safety mechanism that reduces but does not eliminate risk of inappropriate or truncated output. Source: https://docs.aws.amazon.com/polly/latest/dg/generative-voices.html
What ships with it
1 file 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.
- 9d ago First seen · 314 lines · 92 tokens per session scan A c79897311742
amazon-polly is a skill published in the GitHub repository calesthio/generative-media-skills (170 stars, last pushed 2mo ago), licensed MIT. It adds 92 tokens to every session and 5,887 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 2026-09-03.
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-content-extractor
Locates and assembles source passages matching a semantic request. Use for extracting explanations, topics, quotes, demonstrations, or other specifically described content.
cliptalk-interview-editor
Produces a coherent interview edit by combining speaker discovery, topic selection, dialogue context, cleanup, subtitles, and preview. Use for interviews, podcasts, testimonials, or question-and-answer recordings.
cliptalk-shortform-hook-director
Finds and assembles a reviewable short-form cut with a strong opening hook. Use for Shorts, Reels, social clips, talking-head cutdowns, or requests for a punchier opening.
cliptalk-social-reframe-exporter
Creates a review-only 9:16, 4:5, 1:1, or 16:9 version from an existing accepted ClipTalk cut, then checks the rendered preview. Use only when a cut already exists and the user asks to adapt it for Shorts, Reels, Douyin, Xiaohongshu, WeChat Channels, or square feeds; do not use when the user still needs content found…