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 wave-av/mcp-server --skill wave-askgit clone --depth 1 https://github.com/wave-av/mcp-serverWrote 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/wave-av/mcp-server/wave-ask)<a href="https://agentmods.dev/skills/wave-av/mcp-server/wave-ask"><img src="https://agentmods.dev/badge/skills/wave-av/mcp-server/wave-ask.svg" alt="Measured on agentmods" height="20"></a>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 | $0.00041 | $0.01390 |
| Opus 5 | $0.00020 | $0.00695 |
| Sonnet 5 | $0.00008 | $0.00278 |
| Haiku 4.5 | $0.00004 | $0.00139 |
Grade A, and why
wave-ask 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 today.
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 — 101 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skill: wave-ask / wave_compose
wave_composeis the registered successor towave.ask(mcp-server PR4-MCP). It takes the same kind of input and, when aWAVE_API_KEYis configured, also tries the live gatewayPOST /v1/composeroute first (grounding: "gateway"), falling back to exactly the composition logic documented below (grounding: "snapshot") when no key is set or the live call fails, errors, or times out after 3 seconds.wave.askis kept, unchanged, as an offline-only alias for one release — everything below applies to both, exceptwave_compose's input field isintent(notquestion) and its output additionally carriesgrounding.
When to use
An agent has a media-processing goal (captions, clips, dub, identity, x402
rails) and wants the WAVE product/MCP tool/price shape for it before calling.
"Identity" resolves via the real identity_resolve MCP tool only — no
dedicated identity product exists in the catalog yet, so its proposal's
productIds is empty by design, never a fabricated id.
When NOT to use
Already know the route — call it directly. Need the full contract — read
https://api.wave.online/openapi.json or this server's live tools/list
listing instead.
How to call
MCP: wave_compose({ intent: "<goal>", budgetUsd?: <number> }) — registered
alongside every other tool in this package's src/tools/index.ts. (The
deprecated wave.ask({ question: "<goal>", budgetUsd?: <number> }) alias
still works identically for one release.)
{ "intent": "live captions from my mic" }
budgetUsd is optional and never used to compute or invent a price — it only
reorders which suggestion in next[] is surfaced first (a reminder to
confirm the live 402 quote against it before calling).
What you get back
A proposal object:
{
"intent": "live captions from my mic",
"stages": ["realtime", "transcribe", "captions"],
"productIds": ["realtime", "transcribe", "captions"],
"tools": ["perception_subscribe", "wave_create_caption_job", "wave_list_captions", "wave_download_captions"],
"meters": ["wave_realtime_video_minutes", "wave_transcription_minutes", "wave_caption_minutes"],
"priceRows": [
{ "productId": "realtime", "meter": "wave_realtime_video_minutes", "priceShape": "x402 · USDC · base", "quote": "quote at call time" },
{ "productId": "transcribe", "meter": "wave_transcription_minutes", "priceShape": "x402 · USDC · base", "quote": "quote at call time" },
{ "productId": "captions", "meter": "wave_caption_minutes", "priceShape": "x402 · USDC · base", "quote": "quote at call time" }
],
"executes": false,
"next": [
"adjacent capability: translated captions in a second language",
"agent path via MCP: call perception_subscribe directly once you're ready to execute this yourself",
"saved-flow signup: keep this composition for next time (post-GA)"
],
"grounding": "snapshot"
}
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.
- today First seen · 101 lines · 41 tokens per session scan A c1269e88027a
wave-ask is a skill published in the GitHub repository wave-av/mcp-server (1 stars, last pushed today), licensed Apache-2.0. It adds 41 tokens to every session and 1,390 once invoked, about $0.0002 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 2026-09-07.
Other skills, from other repositories
breaking-down-work
Decomposes a goal, feature, epic, or deliverable into a structured task tree (tasks, subtasks, checklists) with dependencies in Astravue. Use when the user wants to break down, decompose, or plan out work, turn an epic or feature into tasks, "split this into tasks", scope a deliverable, or build a work breakdown…
capturing-meeting-actions
Extracts action items from meeting notes, transcripts, or a brain dump and creates them as owned, dated tasks in Astravue. Use when the user pastes meeting notes or a transcript, asks to turn notes or standup discussion into tasks, capture action items, or "make tasks out of this".
triaging-overdue-work
Reviews overdue and at-risk tasks in Astravue, assesses why each is slipping, and proposes reschedule, reassign, deprioritize, or close actions before applying them. Use when the user asks what's overdue, behind schedule, slipping, or at risk, wants to catch up on a project, or clean up late work.
mapping-dependencies
Maps task dependencies in an Astravue project, finds the critical path and blocked work, and detects circular dependencies before any rescheduling. Use when the user asks about dependencies, blockers, what's blocking what, the critical path, or wants to understand or fix the sequencing of a project.
generating-status-report
Generates executive project status reports from Astravue data including task progress, budget consumption, and project health. Use when asked for a status update, project summary, executive report, or "how is the project going?".
managing-custom-fields
Creates, configures, and manages custom fields and dropdown options on tasks in Astravue. Use when working with custom fields, adding dropdown options, setting field values on tasks, or asking "what custom fields does this project have?".