[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7AASCE34265FD9913QGXBNZ":3,"skill-same:01M7AASCE34265FD9913QGXBNZ":149},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":42,"commitSha":43,"skillDirectory":44,"forksCount":45,"repositoryHtmlUrl":46,"repositoryHomepageUrl":6,"ownerHtmlUrl":47,"ownerBlogUrl":6,"ownerDescription":48,"updatedAt":49,"install":50,"resources":63},"01M7AASCE34265FD9913QGXBNZ","polly-voice-deployer",null,"Deploys Amazon Polly text-to-speech configurations with production defaults: engine selection (standard vs neural vs long-form vs generative), voice selection by language, SSML processing (phoneme, emphasis, break, prosody, say-as), lexicon management (custom pronunciation via PLS lexicons), speech marks generation for lip-sync and visual media sync, synthesize-speech API for real-time synthesis, start-speech-synthesis-task for asynchronous batch synthesis to S3, output formats (mp3, ogg_vorbis, pcm, json), sample rates (8000\u002F16000\u002F22050\u002F24000), CloudWatch metrics monitoring, and pricing-per-character budgeting. Emits a READY_TO_DEPLOY checklist with verification commands. Use when configuring Polly TTS, synthesizing speech from text, deploying neural TTS voices, creating custom pronunciation lexicons, generating speech marks for lip-sync. Triggers: synthesize speech, polly neural voice, polly ssml, polly lexicon, speech marks, start speech synthesis task, polly engine selection, polly pricing per character.","jhkchan\u002Faws-agent-skills","jhkchan","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F1083902?v=4","01M7A2SVQYP9GQTPTDXP5JKYQ4","Jacky Chan","01M7A2SVS117YWCTZABB75HZR1",4,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: polly-voice-deployer\ndescription: 'Deploys Amazon Polly text-to-speech configurations with production defaults: engine selection (standard vs neural vs long-form vs generative), voice selection by language, SSML processing (phoneme, emphasis, break, prosody, say-as), lexicon management (custom pronunciation via PLS lexicons), speech marks generation for lip-sync and visual media sync, synthesize-speech API for real-time synthesis, start-speech-synthesis-task for asynchronous batch synthesis to S3, output formats (mp3, ogg_vorbis, pcm, json), sample rates (8000\u002F16000\u002F22050\u002F24000), CloudWatch metrics monitoring, and pricing-per-character budgeting. Emits a READY_TO_DEPLOY checklist with verification commands. Use when configuring Polly TTS, synthesizing speech from text, deploying neural TTS voices, creating custom pronunciation lexicons, generating speech marks for lip-sync. Triggers: synthesize speech, polly neural voice, polly ssml, polly lexicon, speech marks, start speech synthesis task, polly engine selection, polly pricing per character.'\nlicense: Apache-2.0\ncompatibility: 'Agent runtime that reads SKILL.md (Claude Code, Cursor, Windsurf, Codex, Gemini). For live deployment: AWS CLI v2 with polly access and s3 PutObject on the output bucket. Works with Terraform aws_polly_lexicon resources and CloudFormation custom resources for Polly configurations.'\nmetadata:\n  domain: aws-cloudops\n  complexity: high\n  requires_llm: 'true'\n  phase: '1'\n  supports_pipeline: 'true'\n  entry_point: 'false'\n  family: AI\u002FML\n  task_type: deploy\n  skill_class: capability\n  lifecycle_status: active\n  verdict_shape: READY_TO_DEPLOY | PREREQUISITES_MISSING\n  version: 0.1.0\n  author: Jacky Chan — AWS Community Builder\n  tags: aws, polly, tts, cloudops, deploy, ai-ml, provisioning, neural-voice, ssml, lexicon, speech-marks\n  dependencies: aws-orchestrator\n  keywords: aws, polly, text to speech, tts, neural voice, cloudops, deploy, provisioning, ssml, lexicon, speech marks, synthesize speech, lip sync, pricing per character\n  when_to_use: Invoke when the user wants to configure Amazon Polly for text-to-speech synthesis, select between standard\u002Fneural\u002Flong-form\u002F generative engines, use SSML tags for prosody control, create custom pronunciation lexicons, generate speech marks for lip-sync or captioning, run real-time synthesis via synthesize-speech, or batch-synthesize long text to S3 via start-speech-synthesis-task. Do NOT invoke for Amazon Transcribe (speech-to-text), Amazon Lex (conversational bots), or AWS Lambda function deployment.\n---\n\n# Polly Voice Deployer\n\nAn AWS CloudOps agent skill that deploys Amazon Polly text-to-speech\nconfigurations with correct engine, voice, format, and pricing\ndefaults. The skill walks the operator through engine selection\n(standard vs neural vs long-form vs generative), voice selection by\nlanguage and locale, SSML processing capabilities and per-engine\nconstraints, lexicon-based custom pronunciation, speech marks for\nlip-sync alignment, the synthesize-speech real-time API vs the\nstart-speech-synthesis-task asynchronous API, output format and\nsample rate selection, CloudWatch monitoring, and pricing-per-\ncharacter budgeting, captures all synthesis parameters, explains why\neach default matters, and emits a READY_TO_DEPLOY checklist with\ncopy-pasteable verification commands.\n\n## Activation keywords\n\nsynthesize speech, polly neural voice, polly SSML, polly lexicon,\nspeech marks, start speech synthesis task, polly engine selection,\npolly pricing per character.\n\n## STRICT output contract\n\nWhen this skill is invoked with a Polly text-to-speech deployment\nrequest (synthesize speech, select an engine, configure SSML,\ncreate a lexicon, generate speech marks, run a batch synthesis to\nS3, or a partial configuration), the agent MUST respond with the\nREADY_TO_DEPLOY checklist defined in the \"Output format\" section\nusing the literal all-caps labels `POLLY_VOICE:`, `VERDICT:`,\n`CHECKLIST:`, and `VERIFICATION_COMMANDS:`. Do NOT preface the\nchecklist with prose, headings, or disclaimers — emit the block as\nthe first lines of the response. This contract is what assertion-\nbased evals and downstream deployment pipelines rely on; deviating\nfrom the literal labels breaks automation silently.\n\nIf any prerequisite is missing, the verdict is `PREREQUISITES_MISSING`\nwith a specific gap citation in the checklist (marked `[✗]`), and\n`READY_TO_DEPLOY` MUST NOT also appear.\n\n## Quick navigation\n\n| Section | When to read |\n|---|---|\n| Prerequisites | Always — verify before deployment |\n| Step 1 — Engine selection (standard vs neural vs long-form vs generative) | Quality and cost decisions |\n| Step 2 — Voice selection by language | Voice and locale matching |\n| Step 3 — SSML processing (phoneme, emphasis, break, prosody) | Markup control |\n| Step 4 — Lexicon management (custom pronunciation) | Pronunciation overrides |\n| Step 5 — Speech marks (lip-sync alignment) | Visual media sync |\n| Step 6 — synthesize-speech API (real-time) | Short-text synthesis |\n| Step 7 — start-speech-synthesis-task (async to S3) | Long-text batch synthesis |\n| Step 8 — Output format and sample rate | Audio encoding |\n| Step 9 — CloudWatch metrics | Monitoring |\n| Step 10 — Pricing per character | Cost budgeting |\n| Step 11 — Recent features | Latest |\n| NEVER do these things | Review before signing off |\n| Output format | The literal checklist template |\n| references\u002Fengine-and-voice-selection.md | Engine + voice detail |\n| references\u002Fssml-and-lexicons.md | SSML + lexicon detail |\n\n## Mindset\n\n**One-line takeaway:** Amazon Polly converts text to speech using\nfour engine types. Neural voices are higher quality but cost roughly\n4x standard per character. SSML break and emphasis tags only work\nwith the standard engine. Lexicons map custom words to phonetic\nequivalents. Speech marks are generated as a separate request and\nreturned as JSON line-delimited timestamps aligned to the audio.\n\nThree misconceptions dominate Polly misdeployment at provisioning\ntime:\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#common-misconceptions-from-mindset).\n> Three Polly misconceptions: neural SSML support, text-length limits, speech-mark delivery.\n\n## Configuration dependency graph (novel heuristic)\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#configuration-dependency-graph-sequencing-notes).\n> How to sequence deployment using the dependency graph.\n\n| Configuration | Hard dependencies (API error without) | Silent failure \u002F immutability | Enables downstream |\n|---|---|---|---|\n| Engine selection | — (any engine can be chosen) | SSML `\u003Cbreak>` and `\u003Cemphasis>` silently ignored by neural\u002Flong-form\u002Fgenerative | voice list (per-engine voices differ) |\n| Voice ID | engine must support the voice | requesting a standard-only voice with neural returns InvalidParameterValue | language and locale |\n| Output format | — | pcm requires 8000\u002F16000 sample rate; mp3\u002Fogg support 22050\u002F24000 | audio encoding |\n| Sample rate | must be valid for output format | requesting 24000 with pcm silently coerces or errors | audio quality |\n| Lexicon upload | lexicon Name must be unique in account+region | lexicon referenced in synthesis but not uploaded → synthesis fails with LexiconNotFound | custom pronunciation |\n| SSML input | must be well-formed XML | `\u003Cbreak>`\u002F`\u003Cemphasis>` silently ignored by neural engine | prosody control |\n| Speech marks | separate synthesize-speech call with OutputFormat=json | requesting speech marks with OutputFormat=mp3 returns no marks | lip-sync \u002F captioning |\n| Async task (S3) | S3 bucket exists and Polly has WriteS3 access | task fails if S3 permissions missing; check task status | batch synthesis |\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#cross-dependency-gotchas).\n> Baseline-miss rows and gotchas: engine\u002Fvoice subsets, engine SSML support, format-rate coupling, separate marks call, per-region lexicons.\n\n## Expert heuristic: engine selection decision tree\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#expert-heuristic-engine-selection-decision-tree).\n> Engine selection balancing SSML needs, voice availability, and cost.\n\n## Expert heuristic: real-time vs async synthesis\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#expert-heuristic-real-time-vs-async-synthesis).\n> Real-time vs async path by text length and latency needs; marks are a separate request.\n\n## Expert heuristic: pricing per character across engines\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#expert-heuristic-pricing-per-character-across-engines).\n> Per-character pricing by engine with cost estimation and billing rules.\n\n## Prerequisites (verify before deployment)\n\nBefore emitting deployment commands, verify these prerequisites. If\nany are missing, the verdict is **PREREQUISITES_MISSING**.\n\n| Prerequisite | Why it matters | How to verify |\n|---|---|---|\n| Input text available | Polly synthesizes from text input | Confirm text source and encoding (UTF-8) |\n| Engine selected | Determines voice availability, SSML support, and cost | Decide: standard, neural, long-form, generative |\n| Voice ID identified | Must be valid for the selected engine | `aws polly describe-voices --engine \u003Cengine>` |\n| Language\u002Flocale matched | Voice must match the input text language | `aws polly describe-voices --language-code en-US` |\n| Output format chosen | mp3, ogg_vorbis, pcm, or json (speech marks) | Decide based on downstream player |\n| Sample rate valid for format | pcm = 8000\u002F16000; mp3\u002Fogg = 22050\u002F24000 | Match format to rate |\n| SSML tags engine-compatible | `\u003Cbreak>`\u002F`\u003Cemphasis>` only work with standard engine | Verify engine if SSML uses these tags |\n| S3 output bucket (async task) | Batch synthesis writes to S3 | `aws s3 ls s3:\u002F\u002F\u003Cbucket>` |\n| IAM permissions | polly:SynthesizeSpeech or polly:StartSpeechSynthesisTask | Verify IAM policy |\n| Lexicon uploaded (if referenced) | Lexicon must exist before synthesis calls it | `aws polly list-lexicons` |\n\nIf any prerequisite is missing, output `VERDICT: PREREQUISITES_MISSING`\nand cite the specific gap.\n\n## Step 1 — Engine selection (standard vs neural vs long-form vs generative)\n\nAmazon Polly offers four TTS engines. Each has different quality,\nvoice availability, SSML support, and pricing.\n\n| Engine | Quality | Voice availability | SSML support | Cost (per 1M chars) |\n|---|---|---|---|---|\n| standard | Good | ALL voices (all languages) | Full SSML (`\u003Cbreak>`, `\u003Cemphasis>`, `\u003Cprosody>`, `\u003Cphoneme>`, `\u003Csay-as>`) | $4.00 |\n| neural | High | Subset of voices | Partial (`\u003Cprosody>`, `\u003Cphonome>`, `\u003Csay-as>`; NOT `\u003Cbreak>`, `\u003Cemphasis>`) | $16.00 |\n| long-form | Highest (narrative) | Small subset (long-form-optimized voices) | Partial (`\u003Cprosody>`; NOT `\u003Cbreak>`, `\u003Cemphasis>`) | $100.00 |\n| generative | Highest (conversational) | Small subset (generative voices) | Partial (`\u003Cprosody>`; NOT `\u003Cbreak>`, `\u003Cemphasis>`) | $120.00 |\n\n**Critical SSML constraint:** the `\u003Cbreak>` and `\u003Cemphasis>` tags\nare ONLY processed by the standard engine. Neural, long-form, and\ngenerative engines silently ignore these tags. If the use case\nrequires precise pause control, the standard engine is mandatory.\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-1-engine-and-voice-discovery-commands).\n> describe-voices listings per engine for voice discovery.\n\n## Step 2 — Voice selection by language\n\nEach Polly voice is identified by a Voice ID (e.g., `Joanna`,\n`Matthew`, `Lupe`). Voices are language-specific. Not all voices\nare available for all engines.\n\n| Language code | Example voices | Neural available |\n|---|---|---|\n| en-US | Joanna, Matthew, Stephen, Kevin, Danielle, Gregory | Yes (Joanna, Matthew, Stephen, Kevin, Danielle, Gregory) |\n| en-GB | Amy, Emma, Brian | Yes (Amy, Emma, Brian) |\n| en-AU | Nicole, Olivia | Yes (Nicole, Olivia) |\n| es-US | Lupe, Miguel, Penelope, Lucia, Diego | Partial (Lupe, Lucia, Diego) |\n| fr-FR | Lea, Remi | Yes (Lea, Remi) |\n| de-DE | Vicki, Daniel | Yes (Vicki, Daniel) |\n| ja-JP | Kazuha, Tomoko | Partial |\n| ko-KR | Seoyeon | Yes |\n| zh-CN | Zhiyu | Yes |\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-2-voice-discovery-by-language-commands).\n> describe-voices by language-code and per-voice neural support check.\n\n**Select a voice that:**\n1. Matches the input text language (language-code).\n2. Is available on the chosen engine.\n3. Has the desired gender and style.\n\n## Step 3 — SSML processing (phoneme, emphasis, break, prosody)\n\nSSML (Speech Synthesis Markup Language) provides fine-grained control\nover pronunciation, pacing, emphasis, and prosody. Polly supports a\nsubset of SSML tags.\n\n| SSML tag | Purpose | Engine support |\n|---|---|---|\n| `\u003Cbreak>` | Insert a pause (e.g., `\u003Cbreak time=\"500ms\"\u002F>`) | Standard ONLY |\n| `\u003Cemphasis>` | Emphasize a word (e.g., `\u003Cemphasis level=\"strong\">important\u003C\u002Femphasis>`) | Standard ONLY |\n| `\u003Cprosody>` | Control rate, pitch, volume (e.g., `\u003Cprosody rate=\"slow\">text\u003C\u002Fprosody>`) | ALL engines |\n| `\u003Cphoneme>` | Phonetic pronunciation via IPA (e.g., `\u003Cphoneme alphabet=\"ipa\" ph=\"ˈtɒmɑtəʊ\">tomato\u003C\u002Fphoneme>`) | Standard + Neural |\n| `\u003Csay-as>` | Interpret text type (e.g., `\u003Csay-as interpret-as=\"date\" format=\"mdy\">01\u002F15\u002F2024\u003C\u002Fsay-as>`) | ALL engines |\n| `\u003Cspeak>` | Root SSML element (wraps all SSML) | ALL engines |\n| `\u003Csub>` | Substitute pronunciation (e.g., `\u003Csub alias=\" Avenue\">Ave\u003C\u002Fsub>`) | Standard + Neural |\n| `\u003Cw>` | Part-of-speech override | Standard + Neural |\n| `\u003Camazon:effect>` | Whisper, dynamical range compression | Standard + Neural |\n\n> Moved to [references\u002Fssml-and-lexicons.md](references\u002Fssml-and-lexicons.md#step-3-ssml-examples-standard-vs-neural).\n> Full-tag SSML for standard engine; prosody\u002Fphoneme-only SSML for neural.\n\n**Critical:** if the SSML uses `\u003Cbreak>` or `\u003Cemphasis>` and the\nengine is neural\u002Flong-form\u002Fgenerative, the tags are silently ignored.\nThe synthesis succeeds but the pauses and emphasis do not appear in\nthe output audio. This is the #1 Polly deployment surprise.\n\n## Step 4 — Lexicon management (custom pronunciation)\n\nLexicons (PLS — Pronunciation Lexicon Specification) allow custom\npronunciation mappings. A lexicon maps a word to its phonetic\nequivalent so that Polly pronounces it correctly every time.\n\n**Lexicon use cases:**\n- Company names, product names, and acronyms.\n- Domain-specific terminology (medical, legal, technical).\n- Regional pronunciation variants.\n\n> Moved to [references\u002Fssml-and-lexicons.md](references\u002Fssml-and-lexicons.md#step-4-lexicon-management-commands).\n> PLS lexicon file, put-lexicon upload, lexicon use in synthesis, list\u002Fget verification, constraints.\n\n## Step 5 — Speech marks (lip-sync alignment)\n\nSpeech marks are timestamp-aligned markers that map text positions\nto audio positions. They are essential for lip-sync animation,\ncaptioning, and highlighting text as audio plays.\n\n**Speech mark types:**\n\n| Type | Description | Use case |\n|---|---|---|\n| `viseme` | Visual mouth-shape codes for animation | 3D avatar lip-sync |\n| `word` | Word-level start\u002Fend timestamps | Text highlighting, caption sync |\n| `sentence` | Sentence boundaries | Caption segmentation |\n| `ssml` | SSML tag timestamps | SSML mark navigation |\n\n> Moved to [references\u002Fworked-examples.md](references\u002Fworked-examples.md#step-5-speech-marks-generation-and-output).\n> Speech-mark synthesis command, line-delimited JSON output, and field semantics.\n\n**Critical:** speech marks are generated as a SEPARATE call from the\naudio. The audio synthesis (OutputFormat=mp3) and the marks\nsynthesis (OutputFormat=json) must use the SAME engine, voice, text,\nsample rate, and lexicons for the timestamps to align correctly.\n\n## Step 6 — synthesize-speech API (real-time)\n\nThe `synthesize-speech` API is the real-time synthesis endpoint. It\nreturns the audio stream directly in the HTTP response. Use for\nshort text (under ~3000 characters) where low latency is required.\n\n> Moved to [references\u002Fworked-examples.md](references\u002Fworked-examples.md#step-6-synthesize-speech-real-time-examples).\n> Plain-text, SSML, and lexicon real-time synthesis invocations.\n\n**API constraints:**\n- Maximum input text length: ~3000 characters (varies by configuration).\n- Returns audio as a binary stream in the response body.\n- Synchronous — the caller waits for synthesis to complete.\n- Latency: typically 200-600ms for neural, 100-300ms for standard.\n\n## Step 7 — start-speech-synthesis-task (async to S3)\n\nThe `start-speech-synthesis-task` API is the asynchronous synthesis\nendpoint for long text. It writes the output audio file to a\nspecified S3 bucket. Use for batch processing, long-form content\n(articles, books), and pre-generation of audio assets.\n\n> Moved to [references\u002Fworked-examples.md](references\u002Fworked-examples.md#step-7-async-synthesis-task-commands).\n> start-speech-synthesis-task with S3 output, status check, and task listing.\n\n**Task lifecycle:**\n- `scheduled` → `queued` → `inProgress` → `completed` | `failed`\n- On completion, the audio file is available at the S3 URI:\n  `s3:\u002F\u002F\u003Cbucket>\u002F\u003Ckey-prefix>\u002F\u003Ctask-id>.mp3`\n\n**S3 IAM permissions required:**\nPolly needs `s3:PutObject` on the output bucket. The service-linked\nrole `PollySynthesisTaskServiceRole` is typically auto-managed.\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-7-verify-async-task-s3-output).\n> S3 listing check for the completed async synthesis output.\n\n## Step 8 — Output format and sample rate\n\nOutput format and sample rate are coupled. Not all combinations are\nvalid.\n\n| Output format | Valid sample rates | Use case |\n|---|---|---|\n| mp3 | 22050, 24000 | General-purpose audio playback |\n| ogg_vorbis | 22050, 24000 | Web audio, streaming |\n| pcm | 8000, 16000 | Telephony, raw audio processing |\n| json | N\u002FA (speech marks only) | Speech marks \u002F lip-sync |\n\n> Moved to [references\u002Fworked-examples.md](references\u002Fworked-examples.md#step-8-output-format-and-sample-rate-examples).\n> Telephony PCM 8000 Hz and high-quality MP3 24000 Hz synthesis.\n\n**Mismatching format and sample rate** (e.g., PCM at 24000 Hz) will\nproduce an error or silently coerce the output.\n\n## Step 9 — CloudWatch metrics\n\nPolly publishes metrics to CloudWatch for monitoring synthesis\nvolume, errors, and latency.\n\n| Metric | Description |\n|---|---|\n| `RequestCharacters` | Total characters processed (for billing estimation) |\n| `ResponseLatency` | Average synthesis latency |\n| `2XXCount` \u002F `4XXCount` \u002F `5XXCount` | HTTP response code counts |\n| `ThrottledCount` | Throttled request count (if exceeding limits) |\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-9-cloudwatch-metrics-commands).\n> get-metric-statistics for RequestCharacters and a character-budget billing alarm.\n\n## Step 10 — Pricing per character\n\nPolly pricing is per-character of input text. SSML tags are NOT\ncounted — only the text content between tags.\n\n| Engine | Price per 1M characters (us-east-1) | Notes |\n|---|---|---|\n| standard | $4.00 | All voices, full SSML |\n| neural | $16.00 | 4x standard; subset of voices |\n| long-form | $100.00 | Audiobook-quality narration |\n| generative | $120.00 | Most expressive conversational voices |\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-10-pricing-estimation-commands).\n> Character-count and monthly cost estimation one-liners.\n\n**Free tier:** 5 million characters per month for standard voices,\n1 million characters per month for neural voices (for the first 12\nmonths).\n\n## Step 11 — Recent features\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#step-11-recent-features).\n> Recent AWS features (2023-2026): generative engine, long-form enhancements, neural expansion, enhanced metrics, SSE-KMS, viseme improvements.\n\n## NEVER do these things\n\n1. **NEVER use `\u003Cbreak>` or `\u003Cemphasis>` SSML tags with neural,\n   long-form, or generative engines.** These tags are ONLY processed\n   by the standard engine. Neural and others silently ignore them.\n   If the use case requires pause control, use standard engine or\n   `\u003Cprosody rate>` for timing adjustment.\n\n2. **NEVER assume speech marks are returned with the audio.** Speech\n   marks are a SEPARATE `synthesize-speech` call with\n   `OutputFormat=json` and `SpeechMarkTypes` specified. The audio\n   and marks are two different requests.\n\n3. **NEVER use `synthesize-speech` for long text (>3000 chars).**\n   The real-time API has a payload limit. For longer text, use\n   `start-speech-synthesis-task` which writes to S3 and handles\n   arbitrarily long input.\n\n4. **NEVER request a sample rate incompatible with the output\n   format.** PCM supports only 8000 and 16000 Hz. MP3 and OGG\n   support 22050 and 24000 Hz. Mismatching produces errors or\n   silent coercion.\n\n5. **NEVER assume all voices are available on all engines.** Neural,\n   long-form, and generative each support a SUBSET of the standard\n   voice catalog. Always verify with `describe-voices --engine`.\n\n6. **NEVER assume lexicons are global.** Lexicons are per-region.\n   A lexicon uploaded in us-east-1 is NOT available in eu-west-1.\n   Upload to each region where needed.\n\n7. **NEVER forget to verify the S3 output for async tasks.** The\n   task may fail due to IAM permissions or bucket policy. Always\n   check `get-speech-synthesis-task` status and verify the S3\n   object exists.\n\n8. **NEVER estimate cost without checking the engine rate.** Neural\n   is 4x standard, long-form is 25x standard, generative is 30x\n   standard. High-volume workloads must budget by engine.\n\n9. **NEVER use PCM output for web playback.** PCM is raw, unheadered\n   audio. Use MP3 or OGG for web and mobile applications. PCM is\n   for telephony and audio processing pipelines.\n\n10. **NEVER reference a lexicon before uploading it.** The synthesis\n    call will fail with `LexiconNotFound`. Always `put-lexicon`\n    before referencing it in `synthesize-speech`.\n\n## Output format\n\n```text\nPOLLY_VOICE: \u003Cvoice-id> (\u003Cengine> engine, \u003Clanguage-code>)\nVERDICT: READY_TO_DEPLOY | PREREQUISITES_MISSING\nCHECKLIST:\n  [✓|✗] Engine: standard | neural | long-form | generative\n  [✓|✗] Voice: \u003Cvoice-id> (\u003Clanguage-code>, \u003Cgender>)\n  [✓|✗] Voice available on engine: PASS | FAIL (voice not supported)\n  [✓|✗] SSML compatibility: \u003Ctags used> — PASS (all tags supported) | FAIL (\u003Cbreak>\u002F\u003Cemphasis> not supported on neural)\n  [✓|✗] Output format: mp3 | ogg_vorbis | pcm | json\n  [✓|✗] Sample rate: \u003Crate> Hz — PASS (valid for format) | FAIL (invalid for format)\n  [✓|✗] Lexicon: \u003Cname> — uploaded | not referenced\n  [✓|✗] Synthesis mode: real-time (synthesize-speech) | async (start-speech-synthesis-task → s3:\u002F\u002F\u003Cbucket>\u002F\u003Ckey>)\n  [✓|✗] Speech marks: \u003Ctypes> (separate request) | not required\n  [✓|✗] Estimated cost: $\u003Camount> (\u003Cchar-count> chars × $\u003Crate>\u002F1M)\n  [✓|✗] Tags: \u003Ckey=value list>\nVERIFICATION_COMMANDS:\n  aws polly describe-voices --engine \u003Cengine> --query 'Voices[?Id==`\u003Cvoice-id>`]'\n  aws polly list-lexicons --query 'Lexicons[?Name==`\u003Clexicon>`]'\n  aws polly get-speech-synthesis-task --task-id \u003Ctask-id>\n```\n\n### Worked example — neural voice real-time synthesis\n\n```text\nPOLLY_VOICE: Joanna (neural engine, en-US)\nVERDICT: READY_TO_DEPLOY\nCHECKLIST:\n  [✓] Engine: neural\n  [✓] Voice: Joanna (en-US, Female)\n  [✓] Voice available on engine: PASS (Joanna supports neural)\n  [✓] SSML compatibility: plain text (no SSML) — PASS\n  [✓] Output format: mp3\n  [✓] Sample rate: 24000 Hz — PASS (valid for mp3)\n  [✓] Lexicon: company-terms — uploaded\n  [✓] Synthesis mode: real-time (synthesize-speech)\n  [✓] Speech marks: not required\n  [✓] Estimated cost: $0.00 (52 chars × $16\u002F1M = $0.000832)\n  [✓] Tags: Environment=production, UseCase=greeting\nVERIFICATION_COMMANDS:\n  aws polly describe-voices --engine neural --query 'Voices[?Id==`Joanna`]'\n  aws polly list-lexicons --query 'Lexicons[?Name==`company-terms`]'\n```\n\n## Error handling\n\n> Moved to [references\u002Ferror-handling.md](references\u002Ferror-handling.md#error-handling).\n> Symptom-by-symptom fixes: missing pauses, S3 access denied, voice\u002Flexicon not found, empty marks, garbled PCM.\n## References (load on demand)\n\n- [advanced-patterns](references\u002Fadvanced-patterns.md) — expert-heuristic deep dives, misconceptions, dependency-graph notes, recent AWS features (2023-2026)\n- [worked-examples](references\u002Fworked-examples.md) — filled-in synthesis examples (speech marks, real-time, async, output formats)\n- [diagnostic-commands](references\u002Fdiagnostic-commands.md) — voice discovery, task verification, CloudWatch, and pricing commands\n- [error-handling](references\u002Ferror-handling.md) — symptom-by-symptom troubleshooting\n- [engine-and-voice-selection](references\u002Fengine-and-voice-selection.md) — engine and voice selection detail (existing)\n- [ssml-and-lexicons](references\u002Fssml-and-lexicons.md) — SSML and lexicon detail plus examples (existing)\n\n## Domain\n\nAWS CloudOps \u002F Amazon Polly Text-to-Speech Configuration & Voice\nSynthesis Deployment.\n\n## AWS documentation\n\n- **Amazon Polly Developer Guide** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fwhat-is.html\n- **Synthesize speech** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002FAPI_SynthesizeSpeech.html\n- **Start speech synthesis task** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002FAPI_StartSpeechSynthesisTask.html\n- **SSML tags** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fssml.html\n- **Lexicons** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fmanaging-lexicons.html\n- **Speech marks** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fspeechmarks.html\n- **Voice list** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fvoicelist.html\n- **Engine types** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fchoosing-the-engine.html\n- **Pricing** — https:\u002F\u002Faws.amazon.com\u002Fpolly\u002Fpricing\u002F\n- **CloudWatch metrics** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fcloudwatch-monitoring.html\n","# Polly Voice Deployer\n\nAn AWS CloudOps agent skill that deploys Amazon Polly text-to-speech\nconfigurations with correct engine, voice, format, and pricing\ndefaults. The skill walks the operator through engine selection\n(standard vs neural vs long-form vs generative), voice selection by\nlanguage and locale, SSML processing capabilities and per-engine\nconstraints, lexicon-based custom pronunciation, speech marks for\nlip-sync alignment, the synthesize-speech real-time API vs the\nstart-speech-synthesis-task asynchronous API, output format and\nsample rate selection, CloudWatch monitoring, and pricing-per-\ncharacter budgeting, captures all synthesis parameters, explains why\neach default matters, and emits a READY_TO_DEPLOY checklist with\ncopy-pasteable verification commands.\n\n## Activation keywords\n\nsynthesize speech, polly neural voice, polly SSML, polly lexicon,\nspeech marks, start speech synthesis task, polly engine selection,\npolly pricing per character.\n\n## STRICT output contract\n\nWhen this skill is invoked with a Polly text-to-speech deployment\nrequest (synthesize speech, select an engine, configure SSML,\ncreate a lexicon, generate speech marks, run a batch synthesis to\nS3, or a partial configuration), the agent MUST respond with the\nREADY_TO_DEPLOY checklist defined in the \"Output format\" section\nusing the literal all-caps labels `POLLY_VOICE:`, `VERDICT:`,\n`CHECKLIST:`, and `VERIFICATION_COMMANDS:`. Do NOT preface the\nchecklist with prose, headings, or disclaimers — emit the block as\nthe first lines of the response. This contract is what assertion-\nbased evals and downstream deployment pipelines rely on; deviating\nfrom the literal labels breaks automation silently.\n\nIf any prerequisite is missing, the verdict is `PREREQUISITES_MISSING`\nwith a specific gap citation in the checklist (marked `[✗]`), and\n`READY_TO_DEPLOY` MUST NOT also appear.\n\n## Quick navigation\n\n| Section | When to read |\n|---|---|\n| Prerequisites | Always — verify before deployment |\n| Step 1 — Engine selection (standard vs neural vs long-form vs generative) | Quality and cost decisions |\n| Step 2 — Voice selection by language | Voice and locale matching |\n| Step 3 — SSML processing (phoneme, emphasis, break, prosody) | Markup control |\n| Step 4 — Lexicon management (custom pronunciation) | Pronunciation overrides |\n| Step 5 — Speech marks (lip-sync alignment) | Visual media sync |\n| Step 6 — synthesize-speech API (real-time) | Short-text synthesis |\n| Step 7 — start-speech-synthesis-task (async to S3) | Long-text batch synthesis |\n| Step 8 — Output format and sample rate | Audio encoding |\n| Step 9 — CloudWatch metrics | Monitoring |\n| Step 10 — Pricing per character | Cost budgeting |\n| Step 11 — Recent features | Latest |\n| NEVER do these things | Review before signing off |\n| Output format | The literal checklist template |\n| references\u002Fengine-and-voice-selection.md | Engine + voice detail |\n| references\u002Fssml-and-lexicons.md | SSML + lexicon detail |\n\n## Mindset\n\n**One-line takeaway:** Amazon Polly converts text to speech using\nfour engine types. Neural voices are higher quality but cost roughly\n4x standard per character. SSML break and emphasis tags only work\nwith the standard engine. Lexicons map custom words to phonetic\nequivalents. Speech marks are generated as a separate request and\nreturned as JSON line-delimited timestamps aligned to the audio.\n\nThree misconceptions dominate Polly misdeployment at provisioning\ntime:\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#common-misconceptions-from-mindset).\n> Three Polly misconceptions: neural SSML support, text-length limits, speech-mark delivery.\n\n## Configuration dependency graph (novel heuristic)\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#configuration-dependency-graph-sequencing-notes).\n> How to sequence deployment using the dependency graph.\n\n| Configuration | Hard dependencies (API error without) | Silent failure \u002F immutability | Enables downstream |\n|---|---|---|---|\n| Engine selection | — (any engine can be chosen) | SSML `\u003Cbreak>` and `\u003Cemphasis>` silently ignored by neural\u002Flong-form\u002Fgenerative | voice list (per-engine voices differ) |\n| Voice ID | engine must support the voice | requesting a standard-only voice with neural returns InvalidParameterValue | language and locale |\n| Output format | — | pcm requires 8000\u002F16000 sample rate; mp3\u002Fogg support 22050\u002F24000 | audio encoding |\n| Sample rate | must be valid for output format | requesting 24000 with pcm silently coerces or errors | audio quality |\n| Lexicon upload | lexicon Name must be unique in account+region | lexicon referenced in synthesis but not uploaded → synthesis fails with LexiconNotFound | custom pronunciation |\n| SSML input | must be well-formed XML | `\u003Cbreak>`\u002F`\u003Cemphasis>` silently ignored by neural engine | prosody control |\n| Speech marks | separate synthesize-speech call with OutputFormat=json | requesting speech marks with OutputFormat=mp3 returns no marks | lip-sync \u002F captioning |\n| Async task (S3) | S3 bucket exists and Polly has WriteS3 access | task fails if S3 permissions missing; check task status | batch synthesis |\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#cross-dependency-gotchas).\n> Baseline-miss rows and gotchas: engine\u002Fvoice subsets, engine SSML support, format-rate coupling, separate marks call, per-region lexicons.\n\n## Expert heuristic: engine selection decision tree\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#expert-heuristic-engine-selection-decision-tree).\n> Engine selection balancing SSML needs, voice availability, and cost.\n\n## Expert heuristic: real-time vs async synthesis\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#expert-heuristic-real-time-vs-async-synthesis).\n> Real-time vs async path by text length and latency needs; marks are a separate request.\n\n## Expert heuristic: pricing per character across engines\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#expert-heuristic-pricing-per-character-across-engines).\n> Per-character pricing by engine with cost estimation and billing rules.\n\n## Prerequisites (verify before deployment)\n\nBefore emitting deployment commands, verify these prerequisites. If\nany are missing, the verdict is **PREREQUISITES_MISSING**.\n\n| Prerequisite | Why it matters | How to verify |\n|---|---|---|\n| Input text available | Polly synthesizes from text input | Confirm text source and encoding (UTF-8) |\n| Engine selected | Determines voice availability, SSML support, and cost | Decide: standard, neural, long-form, generative |\n| Voice ID identified | Must be valid for the selected engine | `aws polly describe-voices --engine \u003Cengine>` |\n| Language\u002Flocale matched | Voice must match the input text language | `aws polly describe-voices --language-code en-US` |\n| Output format chosen | mp3, ogg_vorbis, pcm, or json (speech marks) | Decide based on downstream player |\n| Sample rate valid for format | pcm = 8000\u002F16000; mp3\u002Fogg = 22050\u002F24000 | Match format to rate |\n| SSML tags engine-compatible | `\u003Cbreak>`\u002F`\u003Cemphasis>` only work with standard engine | Verify engine if SSML uses these tags |\n| S3 output bucket (async task) | Batch synthesis writes to S3 | `aws s3 ls s3:\u002F\u002F\u003Cbucket>` |\n| IAM permissions | polly:SynthesizeSpeech or polly:StartSpeechSynthesisTask | Verify IAM policy |\n| Lexicon uploaded (if referenced) | Lexicon must exist before synthesis calls it | `aws polly list-lexicons` |\n\nIf any prerequisite is missing, output `VERDICT: PREREQUISITES_MISSING`\nand cite the specific gap.\n\n## Step 1 — Engine selection (standard vs neural vs long-form vs generative)\n\nAmazon Polly offers four TTS engines. Each has different quality,\nvoice availability, SSML support, and pricing.\n\n| Engine | Quality | Voice availability | SSML support | Cost (per 1M chars) |\n|---|---|---|---|---|\n| standard | Good | ALL voices (all languages) | Full SSML (`\u003Cbreak>`, `\u003Cemphasis>`, `\u003Cprosody>`, `\u003Cphoneme>`, `\u003Csay-as>`) | $4.00 |\n| neural | High | Subset of voices | Partial (`\u003Cprosody>`, `\u003Cphonome>`, `\u003Csay-as>`; NOT `\u003Cbreak>`, `\u003Cemphasis>`) | $16.00 |\n| long-form | Highest (narrative) | Small subset (long-form-optimized voices) | Partial (`\u003Cprosody>`; NOT `\u003Cbreak>`, `\u003Cemphasis>`) | $100.00 |\n| generative | Highest (conversational) | Small subset (generative voices) | Partial (`\u003Cprosody>`; NOT `\u003Cbreak>`, `\u003Cemphasis>`) | $120.00 |\n\n**Critical SSML constraint:** the `\u003Cbreak>` and `\u003Cemphasis>` tags\nare ONLY processed by the standard engine. Neural, long-form, and\ngenerative engines silently ignore these tags. If the use case\nrequires precise pause control, the standard engine is mandatory.\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-1-engine-and-voice-discovery-commands).\n> describe-voices listings per engine for voice discovery.\n\n## Step 2 — Voice selection by language\n\nEach Polly voice is identified by a Voice ID (e.g., `Joanna`,\n`Matthew`, `Lupe`). Voices are language-specific. Not all voices\nare available for all engines.\n\n| Language code | Example voices | Neural available |\n|---|---|---|\n| en-US | Joanna, Matthew, Stephen, Kevin, Danielle, Gregory | Yes (Joanna, Matthew, Stephen, Kevin, Danielle, Gregory) |\n| en-GB | Amy, Emma, Brian | Yes (Amy, Emma, Brian) |\n| en-AU | Nicole, Olivia | Yes (Nicole, Olivia) |\n| es-US | Lupe, Miguel, Penelope, Lucia, Diego | Partial (Lupe, Lucia, Diego) |\n| fr-FR | Lea, Remi | Yes (Lea, Remi) |\n| de-DE | Vicki, Daniel | Yes (Vicki, Daniel) |\n| ja-JP | Kazuha, Tomoko | Partial |\n| ko-KR | Seoyeon | Yes |\n| zh-CN | Zhiyu | Yes |\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-2-voice-discovery-by-language-commands).\n> describe-voices by language-code and per-voice neural support check.\n\n**Select a voice that:**\n1. Matches the input text language (language-code).\n2. Is available on the chosen engine.\n3. Has the desired gender and style.\n\n## Step 3 — SSML processing (phoneme, emphasis, break, prosody)\n\nSSML (Speech Synthesis Markup Language) provides fine-grained control\nover pronunciation, pacing, emphasis, and prosody. Polly supports a\nsubset of SSML tags.\n\n| SSML tag | Purpose | Engine support |\n|---|---|---|\n| `\u003Cbreak>` | Insert a pause (e.g., `\u003Cbreak time=\"500ms\"\u002F>`) | Standard ONLY |\n| `\u003Cemphasis>` | Emphasize a word (e.g., `\u003Cemphasis level=\"strong\">important\u003C\u002Femphasis>`) | Standard ONLY |\n| `\u003Cprosody>` | Control rate, pitch, volume (e.g., `\u003Cprosody rate=\"slow\">text\u003C\u002Fprosody>`) | ALL engines |\n| `\u003Cphoneme>` | Phonetic pronunciation via IPA (e.g., `\u003Cphoneme alphabet=\"ipa\" ph=\"ˈtɒmɑtəʊ\">tomato\u003C\u002Fphoneme>`) | Standard + Neural |\n| `\u003Csay-as>` | Interpret text type (e.g., `\u003Csay-as interpret-as=\"date\" format=\"mdy\">01\u002F15\u002F2024\u003C\u002Fsay-as>`) | ALL engines |\n| `\u003Cspeak>` | Root SSML element (wraps all SSML) | ALL engines |\n| `\u003Csub>` | Substitute pronunciation (e.g., `\u003Csub alias=\" Avenue\">Ave\u003C\u002Fsub>`) | Standard + Neural |\n| `\u003Cw>` | Part-of-speech override | Standard + Neural |\n| `\u003Camazon:effect>` | Whisper, dynamical range compression | Standard + Neural |\n\n> Moved to [references\u002Fssml-and-lexicons.md](references\u002Fssml-and-lexicons.md#step-3-ssml-examples-standard-vs-neural).\n> Full-tag SSML for standard engine; prosody\u002Fphoneme-only SSML for neural.\n\n**Critical:** if the SSML uses `\u003Cbreak>` or `\u003Cemphasis>` and the\nengine is neural\u002Flong-form\u002Fgenerative, the tags are silently ignored.\nThe synthesis succeeds but the pauses and emphasis do not appear in\nthe output audio. This is the #1 Polly deployment surprise.\n\n## Step 4 — Lexicon management (custom pronunciation)\n\nLexicons (PLS — Pronunciation Lexicon Specification) allow custom\npronunciation mappings. A lexicon maps a word to its phonetic\nequivalent so that Polly pronounces it correctly every time.\n\n**Lexicon use cases:**\n- Company names, product names, and acronyms.\n- Domain-specific terminology (medical, legal, technical).\n- Regional pronunciation variants.\n\n> Moved to [references\u002Fssml-and-lexicons.md](references\u002Fssml-and-lexicons.md#step-4-lexicon-management-commands).\n> PLS lexicon file, put-lexicon upload, lexicon use in synthesis, list\u002Fget verification, constraints.\n\n## Step 5 — Speech marks (lip-sync alignment)\n\nSpeech marks are timestamp-aligned markers that map text positions\nto audio positions. They are essential for lip-sync animation,\ncaptioning, and highlighting text as audio plays.\n\n**Speech mark types:**\n\n| Type | Description | Use case |\n|---|---|---|\n| `viseme` | Visual mouth-shape codes for animation | 3D avatar lip-sync |\n| `word` | Word-level start\u002Fend timestamps | Text highlighting, caption sync |\n| `sentence` | Sentence boundaries | Caption segmentation |\n| `ssml` | SSML tag timestamps | SSML mark navigation |\n\n> Moved to [references\u002Fworked-examples.md](references\u002Fworked-examples.md#step-5-speech-marks-generation-and-output).\n> Speech-mark synthesis command, line-delimited JSON output, and field semantics.\n\n**Critical:** speech marks are generated as a SEPARATE call from the\naudio. The audio synthesis (OutputFormat=mp3) and the marks\nsynthesis (OutputFormat=json) must use the SAME engine, voice, text,\nsample rate, and lexicons for the timestamps to align correctly.\n\n## Step 6 — synthesize-speech API (real-time)\n\nThe `synthesize-speech` API is the real-time synthesis endpoint. It\nreturns the audio stream directly in the HTTP response. Use for\nshort text (under ~3000 characters) where low latency is required.\n\n> Moved to [references\u002Fworked-examples.md](references\u002Fworked-examples.md#step-6-synthesize-speech-real-time-examples).\n> Plain-text, SSML, and lexicon real-time synthesis invocations.\n\n**API constraints:**\n- Maximum input text length: ~3000 characters (varies by configuration).\n- Returns audio as a binary stream in the response body.\n- Synchronous — the caller waits for synthesis to complete.\n- Latency: typically 200-600ms for neural, 100-300ms for standard.\n\n## Step 7 — start-speech-synthesis-task (async to S3)\n\nThe `start-speech-synthesis-task` API is the asynchronous synthesis\nendpoint for long text. It writes the output audio file to a\nspecified S3 bucket. Use for batch processing, long-form content\n(articles, books), and pre-generation of audio assets.\n\n> Moved to [references\u002Fworked-examples.md](references\u002Fworked-examples.md#step-7-async-synthesis-task-commands).\n> start-speech-synthesis-task with S3 output, status check, and task listing.\n\n**Task lifecycle:**\n- `scheduled` → `queued` → `inProgress` → `completed` | `failed`\n- On completion, the audio file is available at the S3 URI:\n  `s3:\u002F\u002F\u003Cbucket>\u002F\u003Ckey-prefix>\u002F\u003Ctask-id>.mp3`\n\n**S3 IAM permissions required:**\nPolly needs `s3:PutObject` on the output bucket. The service-linked\nrole `PollySynthesisTaskServiceRole` is typically auto-managed.\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-7-verify-async-task-s3-output).\n> S3 listing check for the completed async synthesis output.\n\n## Step 8 — Output format and sample rate\n\nOutput format and sample rate are coupled. Not all combinations are\nvalid.\n\n| Output format | Valid sample rates | Use case |\n|---|---|---|\n| mp3 | 22050, 24000 | General-purpose audio playback |\n| ogg_vorbis | 22050, 24000 | Web audio, streaming |\n| pcm | 8000, 16000 | Telephony, raw audio processing |\n| json | N\u002FA (speech marks only) | Speech marks \u002F lip-sync |\n\n> Moved to [references\u002Fworked-examples.md](references\u002Fworked-examples.md#step-8-output-format-and-sample-rate-examples).\n> Telephony PCM 8000 Hz and high-quality MP3 24000 Hz synthesis.\n\n**Mismatching format and sample rate** (e.g., PCM at 24000 Hz) will\nproduce an error or silently coerce the output.\n\n## Step 9 — CloudWatch metrics\n\nPolly publishes metrics to CloudWatch for monitoring synthesis\nvolume, errors, and latency.\n\n| Metric | Description |\n|---|---|\n| `RequestCharacters` | Total characters processed (for billing estimation) |\n| `ResponseLatency` | Average synthesis latency |\n| `2XXCount` \u002F `4XXCount` \u002F `5XXCount` | HTTP response code counts |\n| `ThrottledCount` | Throttled request count (if exceeding limits) |\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-9-cloudwatch-metrics-commands).\n> get-metric-statistics for RequestCharacters and a character-budget billing alarm.\n\n## Step 10 — Pricing per character\n\nPolly pricing is per-character of input text. SSML tags are NOT\ncounted — only the text content between tags.\n\n| Engine | Price per 1M characters (us-east-1) | Notes |\n|---|---|---|\n| standard | $4.00 | All voices, full SSML |\n| neural | $16.00 | 4x standard; subset of voices |\n| long-form | $100.00 | Audiobook-quality narration |\n| generative | $120.00 | Most expressive conversational voices |\n\n> Moved to [references\u002Fdiagnostic-commands.md](references\u002Fdiagnostic-commands.md#step-10-pricing-estimation-commands).\n> Character-count and monthly cost estimation one-liners.\n\n**Free tier:** 5 million characters per month for standard voices,\n1 million characters per month for neural voices (for the first 12\nmonths).\n\n## Step 11 — Recent features\n\n> Moved to [references\u002Fadvanced-patterns.md](references\u002Fadvanced-patterns.md#step-11-recent-features).\n> Recent AWS features (2023-2026): generative engine, long-form enhancements, neural expansion, enhanced metrics, SSE-KMS, viseme improvements.\n\n## NEVER do these things\n\n1. **NEVER use `\u003Cbreak>` or `\u003Cemphasis>` SSML tags with neural,\n   long-form, or generative engines.** These tags are ONLY processed\n   by the standard engine. Neural and others silently ignore them.\n   If the use case requires pause control, use standard engine or\n   `\u003Cprosody rate>` for timing adjustment.\n\n2. **NEVER assume speech marks are returned with the audio.** Speech\n   marks are a SEPARATE `synthesize-speech` call with\n   `OutputFormat=json` and `SpeechMarkTypes` specified. The audio\n   and marks are two different requests.\n\n3. **NEVER use `synthesize-speech` for long text (>3000 chars).**\n   The real-time API has a payload limit. For longer text, use\n   `start-speech-synthesis-task` which writes to S3 and handles\n   arbitrarily long input.\n\n4. **NEVER request a sample rate incompatible with the output\n   format.** PCM supports only 8000 and 16000 Hz. MP3 and OGG\n   support 22050 and 24000 Hz. Mismatching produces errors or\n   silent coercion.\n\n5. **NEVER assume all voices are available on all engines.** Neural,\n   long-form, and generative each support a SUBSET of the standard\n   voice catalog. Always verify with `describe-voices --engine`.\n\n6. **NEVER assume lexicons are global.** Lexicons are per-region.\n   A lexicon uploaded in us-east-1 is NOT available in eu-west-1.\n   Upload to each region where needed.\n\n7. **NEVER forget to verify the S3 output for async tasks.** The\n   task may fail due to IAM permissions or bucket policy. Always\n   check `get-speech-synthesis-task` status and verify the S3\n   object exists.\n\n8. **NEVER estimate cost without checking the engine rate.** Neural\n   is 4x standard, long-form is 25x standard, generative is 30x\n   standard. High-volume workloads must budget by engine.\n\n9. **NEVER use PCM output for web playback.** PCM is raw, unheadered\n   audio. Use MP3 or OGG for web and mobile applications. PCM is\n   for telephony and audio processing pipelines.\n\n10. **NEVER reference a lexicon before uploading it.** The synthesis\n    call will fail with `LexiconNotFound`. Always `put-lexicon`\n    before referencing it in `synthesize-speech`.\n\n## Output format\n\n```text\nPOLLY_VOICE: \u003Cvoice-id> (\u003Cengine> engine, \u003Clanguage-code>)\nVERDICT: READY_TO_DEPLOY | PREREQUISITES_MISSING\nCHECKLIST:\n  [✓|✗] Engine: standard | neural | long-form | generative\n  [✓|✗] Voice: \u003Cvoice-id> (\u003Clanguage-code>, \u003Cgender>)\n  [✓|✗] Voice available on engine: PASS | FAIL (voice not supported)\n  [✓|✗] SSML compatibility: \u003Ctags used> — PASS (all tags supported) | FAIL (\u003Cbreak>\u002F\u003Cemphasis> not supported on neural)\n  [✓|✗] Output format: mp3 | ogg_vorbis | pcm | json\n  [✓|✗] Sample rate: \u003Crate> Hz — PASS (valid for format) | FAIL (invalid for format)\n  [✓|✗] Lexicon: \u003Cname> — uploaded | not referenced\n  [✓|✗] Synthesis mode: real-time (synthesize-speech) | async (start-speech-synthesis-task → s3:\u002F\u002F\u003Cbucket>\u002F\u003Ckey>)\n  [✓|✗] Speech marks: \u003Ctypes> (separate request) | not required\n  [✓|✗] Estimated cost: $\u003Camount> (\u003Cchar-count> chars × $\u003Crate>\u002F1M)\n  [✓|✗] Tags: \u003Ckey=value list>\nVERIFICATION_COMMANDS:\n  aws polly describe-voices --engine \u003Cengine> --query 'Voices[?Id==`\u003Cvoice-id>`]'\n  aws polly list-lexicons --query 'Lexicons[?Name==`\u003Clexicon>`]'\n  aws polly get-speech-synthesis-task --task-id \u003Ctask-id>\n```\n\n### Worked example — neural voice real-time synthesis\n\n```text\nPOLLY_VOICE: Joanna (neural engine, en-US)\nVERDICT: READY_TO_DEPLOY\nCHECKLIST:\n  [✓] Engine: neural\n  [✓] Voice: Joanna (en-US, Female)\n  [✓] Voice available on engine: PASS (Joanna supports neural)\n  [✓] SSML compatibility: plain text (no SSML) — PASS\n  [✓] Output format: mp3\n  [✓] Sample rate: 24000 Hz — PASS (valid for mp3)\n  [✓] Lexicon: company-terms — uploaded\n  [✓] Synthesis mode: real-time (synthesize-speech)\n  [✓] Speech marks: not required\n  [✓] Estimated cost: $0.00 (52 chars × $16\u002F1M = $0.000832)\n  [✓] Tags: Environment=production, UseCase=greeting\nVERIFICATION_COMMANDS:\n  aws polly describe-voices --engine neural --query 'Voices[?Id==`Joanna`]'\n  aws polly list-lexicons --query 'Lexicons[?Name==`company-terms`]'\n```\n\n## Error handling\n\n> Moved to [references\u002Ferror-handling.md](references\u002Ferror-handling.md#error-handling).\n> Symptom-by-symptom fixes: missing pauses, S3 access denied, voice\u002Flexicon not found, empty marks, garbled PCM.\n## References (load on demand)\n\n- [advanced-patterns](references\u002Fadvanced-patterns.md) — expert-heuristic deep dives, misconceptions, dependency-graph notes, recent AWS features (2023-2026)\n- [worked-examples](references\u002Fworked-examples.md) — filled-in synthesis examples (speech marks, real-time, async, output formats)\n- [diagnostic-commands](references\u002Fdiagnostic-commands.md) — voice discovery, task verification, CloudWatch, and pricing commands\n- [error-handling](references\u002Ferror-handling.md) — symptom-by-symptom troubleshooting\n- [engine-and-voice-selection](references\u002Fengine-and-voice-selection.md) — engine and voice selection detail (existing)\n- [ssml-and-lexicons](references\u002Fssml-and-lexicons.md) — SSML and lexicon detail plus examples (existing)\n\n## Domain\n\nAWS CloudOps \u002F Amazon Polly Text-to-Speech Configuration & Voice\nSynthesis Deployment.\n\n## AWS documentation\n\n- **Amazon Polly Developer Guide** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fwhat-is.html\n- **Synthesize speech** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002FAPI_SynthesizeSpeech.html\n- **Start speech synthesis task** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002FAPI_StartSpeechSynthesisTask.html\n- **SSML tags** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fssml.html\n- **Lexicons** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fmanaging-lexicons.html\n- **Speech marks** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fspeechmarks.html\n- **Voice list** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fvoicelist.html\n- **Engine types** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fchoosing-the-engine.html\n- **Pricing** — https:\u002F\u002Faws.amazon.com\u002Fpolly\u002Fpricing\u002F\n- **CloudWatch metrics** — https:\u002F\u002Fdocs.aws.amazon.com\u002Fpolly\u002Flatest\u002Fdg\u002Fcloudwatch-monitoring.html\n",{"name":5,"license":15,"metadata":24,"description":7,"compatibility":41},{"tags":25,"phase":26,"author":27,"domain":28,"family":29,"version":30,"keywords":31,"task_type":32,"complexity":33,"entry_point":34,"skill_class":35,"when_to_use":36,"dependencies":37,"requires_llm":38,"verdict_shape":39,"lifecycle_status":40,"supports_pipeline":38},"aws, polly, tts, cloudops, deploy, ai-ml, provisioning, neural-voice, ssml, lexicon, speech-marks","1","Jacky Chan — AWS Community Builder","aws-cloudops","AI\u002FML","0.1.0","aws, polly, text to speech, tts, neural voice, cloudops, deploy, provisioning, ssml, lexicon, speech marks, synthesize speech, lip sync, pricing per character","deploy","high","false","capability","Invoke when the user wants to configure Amazon Polly for text-to-speech synthesis, select between standard\u002Fneural\u002Flong-form\u002F generative engines, use SSML tags for prosody control, create custom pronunciation lexicons, generate speech marks for lip-sync or captioning, run real-time synthesis via synthesize-speech, or batch-synthesize long text to S3 via start-speech-synthesis-task. Do NOT invoke for Amazon Transcribe (speech-to-text), Amazon Lex (conversational bots), or AWS Lambda function deployment.","aws-orchestrator","true","READY_TO_DEPLOY | PREREQUISITES_MISSING","active","Agent runtime that reads SKILL.md (Claude Code, Cursor, Windsurf, Codex, Gemini). For live deployment: AWS CLI v2 with polly access and s3 PutObject on the output bucket. Works with Terraform aws_polly_lexicon resources and CloudFormation custom resources for Polly configurations.","main","eeace51bd5dd1db2030729d83eeab988861553aa","skills\u002Fpolly-voice-deployer",0,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills","https:\u002F\u002Fgithub.com\u002Fjhkchan","CTO & Co-Founder @Beever-AI  & @votee. CAISP. AWS Community Builder. Organizer@GDGCloudHK. MongoDB Champion. Technopreneur + Infornographer.","2026-09-24 08:54:10",{"source":8,"skillName":5,"commands":51,"prompt":61,"zipAvailable":62},[52,55,58],{"tool":53,"command":54},"npx","npx skills add jhkchan\u002Faws-agent-skills --skill polly-voice-deployer",{"tool":56,"command":57},"bunx","bunx skills add jhkchan\u002Faws-agent-skills --skill polly-voice-deployer",{"tool":59,"command":60},"pnpm","pnpm dlx skills add jhkchan\u002Faws-agent-skills --skill polly-voice-deployer","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002FSKILL.md",true,[64,70,74,78,82,86,90,94,98,102,106,109,113,117,121,125,129,133,137,141,145],{"path":65,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":68,"sourceHtmlUrl":69},"eval\u002Ftest-cases.yaml","REFERENCE","参考文档",5836,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Feval\u002Ftest-cases.yaml",{"path":71,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":72,"sourceHtmlUrl":73},"evals\u002Fbaselines\u002Fasync-s3-batch.md",804,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fbaselines\u002Fasync-s3-batch.md",{"path":75,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":76,"sourceHtmlUrl":77},"evals\u002Fbaselines\u002Fneural-realtime-lexicon.md",703,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fbaselines\u002Fneural-realtime-lexicon.md",{"path":79,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":80,"sourceHtmlUrl":81},"evals\u002Fbaselines\u002Fspeech-marks-lipsync.md",872,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fbaselines\u002Fspeech-marks-lipsync.md",{"path":83,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":84,"sourceHtmlUrl":85},"evals\u002Fbaselines\u002Fssml-break-on-neural.md",819,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fbaselines\u002Fssml-break-on-neural.md",{"path":87,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":88,"sourceHtmlUrl":89},"evals\u002Fbaselines\u002Fssml-standard-engine.md",861,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fbaselines\u002Fssml-standard-engine.md",{"path":91,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":92,"sourceHtmlUrl":93},"evals\u002Feval_config.yaml",1704,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Feval_config.yaml",{"path":95,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":96,"sourceHtmlUrl":97},"evals\u002Fevals.json",9294,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fevals.json",{"path":99,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":100,"sourceHtmlUrl":101},"evals\u002Fprompts\u002Fasync-s3-batch.md",523,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fprompts\u002Fasync-s3-batch.md",{"path":103,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":104,"sourceHtmlUrl":105},"evals\u002Fprompts\u002Fneural-realtime-lexicon.md",668,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fprompts\u002Fneural-realtime-lexicon.md",{"path":107,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":100,"sourceHtmlUrl":108},"evals\u002Fprompts\u002Fspeech-marks-lipsync.md","https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fprompts\u002Fspeech-marks-lipsync.md",{"path":110,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":111,"sourceHtmlUrl":112},"evals\u002Fprompts\u002Fssml-break-on-neural.md",456,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fprompts\u002Fssml-break-on-neural.md",{"path":114,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":115,"sourceHtmlUrl":116},"evals\u002Fprompts\u002Fssml-standard-engine.md",527,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fevals\u002Fprompts\u002Fssml-standard-engine.md",{"path":118,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":119,"sourceHtmlUrl":120},"examples\u002FREADME.md",5552,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fexamples\u002FREADME.md",{"path":122,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":123,"sourceHtmlUrl":124},"references\u002Fadvanced-patterns.md",7409,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Freferences\u002Fadvanced-patterns.md",{"path":126,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":127,"sourceHtmlUrl":128},"references\u002Fdiagnostic-commands.md",2089,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Freferences\u002Fdiagnostic-commands.md",{"path":130,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":131,"sourceHtmlUrl":132},"references\u002Fengine-and-voice-selection.md",7625,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Freferences\u002Fengine-and-voice-selection.md",{"path":134,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":135,"sourceHtmlUrl":136},"references\u002Ferror-handling.md",1354,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Freferences\u002Ferror-handling.md",{"path":138,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":139,"sourceHtmlUrl":140},"references\u002Fssml-and-lexicons.md",11995,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Freferences\u002Fssml-and-lexicons.md",{"path":142,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":143,"sourceHtmlUrl":144},"references\u002Fworked-examples.md",3045,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Freferences\u002Fworked-examples.md",{"path":146,"resourceType":66,"resourceTypeDesc":67,"sizeBytes":147,"sourceHtmlUrl":148},"skill.usf.yaml",4570,"https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fpolly-voice-deployer\u002Fskill.usf.yaml",[150,157,164,171,178,185],{"skillNo":151,"name":152,"nameZh":6,"description":153,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":154,"categories":155,"keywords":156,"publishedAt":19},"01M7ADVD60X4PCAPKSE58P9NDC","xray-tracing-deployer","Deploys AWS X-Ray distributed tracing with production-grade configuration: X-Ray daemon deployment (EC2 systemd, ECS sidecar, EKS DaemonSet, Lambda built-in), SDK instrumentation (Java, Python, Node.js, Go, .NET) with annotations and metadata, sampling rules (default reservoir + rate vs custom rules with URL \u002F method \u002F service-name predicates), service map generation, trace query and retrieval, groups and insights, CloudWatch ServiceLens integration, IAM permissions (PutTraceSegments, GetSamplingRules), and latest features (Lambda Powertools tracing, OpenTelemetry \u002F ADOT Collector support, W3C trace context). Emits a READY_TO_DEPLOY checklist with every configuration item verified. Use when enabling X-Ray on a new application, adding tracing to ECS \u002F EKS \u002F Lambda, validating a sampling configuration, or generating daemon deployment CLI and IaC templates. Triggers: X-Ray, distributed tracing, X-Ray daemon, sampling rules, service map, X-Ray SDK, OpenTelemetry, ADOT, Lambda Powertools, CloudWatch ServiceLens.","https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fxray-tracing-deployer\u002FSKILL.md",[],[],{"skillNo":158,"name":159,"nameZh":6,"description":160,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":161,"categories":162,"keywords":163,"publishedAt":19},"01M7ADTVJ7DCBM9CB1FY51GH7J","wellarchitected-workload-auditor","Audits AWS Well-Architected Tool workloads for review staleness, per-pillar high-risk issue counts, milestone tracking gaps, and remediation plan completeness. Emits a deterministic verdict (STALE_REVIEW | HIGH_RISK | CONFIG_GAP | OK) per workload with per-pillar risk breakdown, milestone trend analysis, and specific remediation. Use when reviewing Well-Architected workloads, checking WA review freshness, auditing high-risk issues per pillar, validating milestone coverage, or tracking remediation plans.","https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fwellarchitected-workload-auditor\u002FSKILL.md",[],[],{"skillNo":165,"name":166,"nameZh":6,"description":167,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":168,"categories":169,"keywords":170,"publishedAt":19},"01M7ADT59G6FTTNZYHT2QRPNNP","wellarchitected-review-operator","Operates AWS Well-Architected Tool reviews end-to-end — creates workloads, runs pillar reviews (operational excellence, security, reliability, performance, cost optimization, sustainability, Prosperity), answers questions with risk-tier rationale, generates reports and improvement plans, creates milestones, and integrates with Trusted Advisor and Well-Architected Labs. Runs deterministic pre-checks (workload existence, lens availability, pillar coverage), emits the exact update-answer \u002F get-consolidated-report CLI behind a CONFIRM gate, and verifies risk-tier transitions. Emits READY | BLOCKED | COMPLETED. Use when creating a workload, answering pillar questions, generating an improvement plan, creating a milestone, or integrating with Trusted Advisor. Triggers: well-architected tool, well architected review, workload review, pillar, operational excellence, security, reliability, performance, cost optimization, sustainability, prosperity, improvement plan, milestone, trusted advisor.","https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fwellarchitected-review-operator\u002FSKILL.md",[],[],{"skillNo":172,"name":173,"nameZh":6,"description":174,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":175,"categories":176,"keywords":177,"publishedAt":19},"01M7ADSH5JA9QHNPPHQCD7EVA7","wafv2-web-acl-deployer","Provisions AWS WAFv2 Web ACLs with correct production defaults: scope selection (CLOUDFRONT vs REGIONAL), managed rule groups (AWSManagedRulesCommonRuleSet, SQLiRuleSet, LinuxRuleSet, WindowsRuleSet, IPReputationList, BotControlRuleSet, AmazonIpReputationList, ATP), custom rules (byte match, IP set, regex pattern, geo match, size constraint, rate-based), rule priority ordering, logging destinations (Kinesis Firehose, CloudWatch Logs, S3), resource association (ALB, API Gateway, CloudFront), rate-based rules with aggregate keys (IP, forwarded IP), Challenge and CAPTCHA actions, and account takeover prevention. Emits a READY_TO_DEPLOY checklist. Use when creating a Web ACL, attaching managed rules, configuring rate limiting, enabling WAF logging, associating a Web ACL with an ALB\u002FAPI Gateway\u002FCloudFront, or configuring CAPTCHA\u002FChallenge actions. Triggers: create WAF Web ACL, WAFv2, managed rule groups, WAF rate limiting, WAF logging, WAF CAPTCHA, WAF Challenge action, bot control, ATP managed rules.","https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fwafv2-web-acl-deployer\u002FSKILL.md",[],[],{"skillNo":179,"name":180,"nameZh":6,"description":181,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":182,"categories":183,"keywords":184,"publishedAt":19},"01M7ADRYFAT8R73FBNCHRC573X","wafv2-web-acl-auditor","Audits AWS WAFv2 Web ACL configurations to determine whether the ACL provides effective protection: default-action posture (Allow vs Block), managed-rule-group coverage gaps, rule effectiveness (BLOCK vs COUNT, shadow\u002Fbypass rules, stale exclusions), rate-based rule correctness, logging and visibility configuration, and text-transformation bypass vectors. Emits a deterministic verdict per Web ACL. Use when reviewing a WAFv2 Web ACL for security posture, checking managed-rule coverage, validating rate-limit configuration, auditing WAF logging, or hardening a WAF before production deployment. Triggers: WAF, WAFv2, Web ACL, managed rule group, AWSManagedRules, rate-based rule, CAPTCHA, Challenge, OverrideAction, Count mode, forwarded IP, Bot Control, ATP, text transformation, WAF logging, WAF visibility, OWASP, SQLi, XSS, RFI.","https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fwafv2-web-acl-auditor\u002FSKILL.md",[],[],{"skillNo":186,"name":187,"nameZh":6,"description":188,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":189,"categories":190,"keywords":191,"publishedAt":19},"01M7ADR8W7G72JX4KFGN03QK8Q","waf-rule-deployer","Provisions AWS WAFv2 rule sets with production defaults: Web ACL creation (CloudFront scope in us-east-1 vs Regional scope), rule groups (managed vs custom) with capacity planning in WCU (Web ACL Capacity Units, 1500 default per ACL), managed rule groups (AWSManagedRulesCommonRuleSet, AWSManagedRulesSQLiRuleSet, AWSManagedRulesAmazonIpReputationList, AWSManagedRulesLinuxRuleSet, AWSManagedRulesWindowsRuleSet, AWSManagedRulesUnixRuleSet), custom rules (byte-match, regex pattern set, geo-match, rate-based), IP set and regex pattern set resources, rule priority ordering (lowest number evaluated first), action types (allow, block, count, CAPTCHA, Challenge), bot control managed rule group, account takeover. Triggers: create WAF rule, WAF managed rule group, WAF custom rule, WAF byte match, WAF geo match, WAF rate based rule, WAF IP set, WAF regex pattern set, WAF WCU budget, WAF CloudFront ACL, WAF logging Kinesis Firehose, WAF bot control, WAF ATP, WAF label matching, WAF Challenge action, WAF CAPTCHA.","https:\u002F\u002Fgithub.com\u002Fjhkchan\u002Faws-agent-skills\u002Fblob\u002Fmain\u002Fskills\u002Fwaf-rule-deployer\u002FSKILL.md",[],[]]