← 返回 Skills 市场
badajoz95

H-ear

作者 Paul Day · GitHub ↗ · v1.1.7 · MIT-0
cross-platform ⚠ suspicious
522
总下载
1
收藏
0
当前安装
33
版本数
在 OpenClaw 中安装
/install h-ear
功能描述
H-ear.world transforms sound into an actionable, meaningful translation layer of the world around you. Describe, share and act upon environmental audio as a...
使用说明 (SKILL.md)

\x3C!-- Version: keep in sync with package.json. Runtime code reads from package.json via createRequire — this header is the static manifest the OpenClaw gateway parses at install/registration time. Bump both together. -->

H-ear — Sound Intelligence for AI Agents

H-ear.world transforms sound into an actionable, meaningful translation layer of the world around you. Describe, share and act upon environmental audio as a machine sensor that empowers you, your business and your AI flow.

Commands

\x3C!-- @doc-sync-ref gen:#openclaw-commands -->

Command Description
h-ear health Check H-ear API health and liveness (no auth required).
h-ear usage Show API usage statistics (minutes, calls, quota).
h-ear sounds \x3Csearch> List supported sound classes (521+ across 3 taxonomies).
h-ear jobs List recent classification jobs (paginated).
h-ear job \x3CjobId> Get detail for one job: status, fileName, eventCount, timestamps.
h-ear job events \x3CjobId> Get noise events for a completed job (timeline order, filterable).
h-ear job audio \x3CjobId> Get a 1-hour SAS URL to stream the source audio of a job.
h-ear job waveform \x3CjobId> Get pre-computed peaks.js waveform + audio URL.
h-ear job report \x3CjobId> Get a 7-day SAS URL to download the Excel analysis report.
h-ear classify \x3CfileOrUrl> Classify audio (URL or local file). Polls until complete.
h-ear webhook list List enterprise webhook registrations.
h-ear webhook get \x3CwebhookId> Show one webhook (URL, events, filter config, delivery stats).
h-ear webhook create \x3Curl> Create an enterprise webhook (returns signing secret ONCE).
h-ear webhook update \x3CwebhookId> Update webhook URL, status (active/paused), or filters.
h-ear webhook delete \x3CwebhookId> Permanently delete a webhook (cannot be undone).
h-ear webhook ping \x3CwebhookId> Send a test ping to a webhook (verify connectivity + signing).
h-ear webhook deliveries \x3CwebhookId> Audit trail: recent delivery attempts for a webhook.
\x3C!-- @end-doc-ref -->

For per-command help with all flags and examples, run h-ear \x3Ccommand> --help. For machine-readable schemas (LLM-friendly introspection): h-ear --list-tools --json.

Setup

Variable Required Default Description
HEAR_API_KEY Yes* H-ear Enterprise API key (ncm_sk_...). Required unless HEAR_BEARER_TOKEN is set. Get one at h-ear.world.
HEAR_BEARER_TOKEN Yes* OAuth bearer token. Alternative to HEAR_API_KEY — one of the two must be set.
HEAR_ENV Yes Target environment: dev, staging, or prod.
HEAR_BASE_URL No Per-environment default Override API base URL (advanced).

*One of HEAR_API_KEY or HEAR_BEARER_TOKEN is required.

Webhook Delivery

Batch classification (classify batch) and sound alerts (alerts on) use webhook callbacks for asynchronous result delivery. The OpenClaw gateway manages webhook endpoints automatically -- the skill registers callbacks against the gateway's own webhook receiver, which routes results back to your connected messaging channel. No external endpoint configuration is required by the user.

Webhook events: job.completed, job.failed, batch.completed, quota.warning.

安全使用建议
This skill mostly does what it says — audio classification and job/webhook management — but there are multiple inconsistencies you should consider before installing: - ffmpeg dependency: The code can capture RTSP streams by spawning 'ffmpeg', but the skill's metadata does not declare ffmpeg as a required binary. If you plan to use the 'capture'/'listen' commands, expect the skill to run ffmpeg and write temporary WAV files. - Undeclared environment variables: The code accepts HEAR_BEARER_TOKEN, HEAR_BASE_URL, and LISTEN_RTSP_URL in addition to HEAR_API_KEY and HEAR_ENV. HEAR_BEARER_TOKEN is effectively required as an alternative to HEAR_API_KEY but wasn't listed in the registry's required env list. If you provide environment variables, be sure you understand which keys/tokens you are granting to the skill. - Webhook behavior mismatch: SKILL.md claims the OpenClaw gateway will manage webhook endpoints automatically, but the code often expects an explicit callbackUrl (alerts registration throws if none). Clarify whether the gateway will inject callback URLs or you must provide them; otherwise webhook registration may fail or require you to supply external URLs that receive events. - Local file and RTSP access: The skill will read arbitrary local files you pass to classify, and capture from RTSP sources if used. Only provide files/RTSP URLs you trust; the skill will upload audio content to the H-ear service as part of classification. If you still want to install: ask the publisher (or check upstream repo) to update the SKILL.md/registry metadata to declare ffmpeg as an optional required binary and to list HEAR_BEARER_TOKEN, HEAR_BASE_URL, and LISTEN_RTSP_URL where they are used. Also confirm the actual webhook integration model (gateway-managed vs user-provided callback URLs). These fixes would resolve the main incoherences noted above.
能力标签
requires-oauth-tokenrequires-sensitive-credentials
能力评估
Purpose & Capability
The skill's name, description, and commands (classify, jobs, webhooks, etc.) align with an audio-classification service. However, the code contains an RTSP capture routine that invokes ffmpeg (spawnSync('ffmpeg')) and references LISTEN_RTSP_URL — yet the registry metadata lists no required binaries. Capture functionality and ffmpeg dependency are not declared, which is inconsistent with the stated registry requirements.
Instruction Scope
SKILL.md states the gateway manages webhooks automatically and that no external endpoint configuration is required. The code, however, frequently accepts/requires explicit callbackUrl values (alerts.ts throws if no callbackUrl) and many classify/webhook functions accept callbackSecret/callbackUrl options. The CLI also reads local files for classification and will spawn ffmpeg for RTSP capture. The instructions therefore promise a hands-off webhook experience but the code expects or demands explicit callback URLs in places — a contradiction that affects what data is produced/transmitted and how webhooks are configured.
Install Mechanism
There is no install spec (instruction-only), which reduces install-time risk. But the package includes full source files and a package.json with dependencies (notably @h-ear/core). That source contains runtime exec of ffmpeg and file I/O. The absence of an install script is not inherently unsafe, but presence of executable child_process calls and file manipulation means the skill will perform non-trivial local I/O if invoked.
Credentials
Registry metadata declares HEAR_API_KEY and HEAR_ENV as required (primary is HEAR_API_KEY). The runtime/README and code also use HEAR_BEARER_TOKEN (alternative auth), HEAR_BASE_URL, and LISTEN_RTSP_URL, but these are not listed in the skill's required env vars. The code will throw if neither HEAR_API_KEY nor HEAR_BEARER_TOKEN is set, so HEAR_BEARER_TOKEN should be declared. LISTEN_RTSP_URL and HEAR_BASE_URL are used but undeclared. This mismatch makes it unclear what secrets/environment the skill needs and what it might read from the environment.
Persistence & Privilege
always is false and the skill does not request persistent platform privileges. It does not appear to modify other skills or global agent configuration. It will, however, read local files (for classify <file>) and create temporary files during RTSP capture.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install h-ear
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /h-ear 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.1.7
- Major update: Expanded CLI with new audio job and webhook commands for improved sound intelligence workflows. - Added new commands: `job events`, `job audio`, `job waveform`, `job report`, and detailed webhook management (list, get, create, update, delete, ping, deliveries). - Command naming standardized: All commands now consistently use `h-ear` prefix (e.g., `h-ear classify`, `h-ear webhook list`). - Enhanced documentation in SKILL.md with updated command details and usage instructions. - Internal refactoring: Added manifest, updated CLI and formatter logic to support new features. - Improved per-command help and introspection options (`--help`, `--list-tools --json`).
v1.1.6
- Improved handling or formatting of job event outputs. - Internal updates in src/commands/job-events.ts and src/formatter.ts for better maintainability.
v1.1.5
Version 1.1.5 of h-ear - No file changes detected in this release. - No updates to commands, setup, or documentation. - Version update only; functionality remains unchanged.
v1.0.15
No user-facing changes detected in this version. - No file changes between versions 0.1.0 and 1.0.15. - No updates to documentation, commands, or environment requirements.
v1.0.14
Version 1.0.14 — No code or documentation changes detected. - No file or feature changes in this release. - Version increment only; functionality and documentation remain the same.
v1.1.2
- Updated description to clarify that audio is treated as a machine sensor and specifically refers to environmental audio. - No changes to commands, setup, or webhook documentation.
v1.1.1
No user-visible changes in this release. - Version incremented to 1.1.1 with no detected file or documentation changes.
v1.1.0
No visible changes detected in this version. - Version number change only; all features and configurations remain the same.
v1.0.10
Version 1.0.10 - No file changes detected for this release. - No updates or new features documented.
v1.0.9
No changes detected in this version. - No file or documentation updates. - Functionality and interface remain unchanged.
v1.0.8
Version 1.0.8 - No file changes detected in this release. - Functionality and documentation remain unchanged from the previous version.
v1.0.7
No user-facing changes in this release.
v1.0.6
Version 1.0.6 - No file changes detected in this release. - No updates to documentation or commands. - Functionality and setup remain unchanged from the previous version.
v1.0.5
Version 1.0.5 of h-ear - No file changes detected in this version. - Functionality and documentation remain unchanged from the previous release.
v1.0.4
- Version bump to 1.0.4 with no file or documentation changes. - No feature updates, bug fixes, or internal modifications detected.
v1.0.3
Version 1.0.3 - No changes detected in this release. - All documentation and configurations remain unchanged from the previous version.
v1.0.2
Version 1.0.2 of h-ear - No changes detected since the previous version. - All features, commands, and documentation remain unchanged.
v1.0.1
- Updated package.json to bump version from 1.0.0 to 1.0.1. - No changes to commands, documentation, or features. - Maintenance release.
v1.0.0
- Initial release of h-ear skill, version 1.0.0. - Provides audio classification, batch processing, sound class listing, usage statistics, job management, real-time alerts, and API health check commands. - Supports both API key and bearer token authentication. - Includes automated webhook integration for asynchronous results and alerts. - No external webhook setup required; events are delivered via the gateway.
v0.2.3
No user-facing changes in this release. - Version incremented to 0.2.3. - No updates to documentation or commands.
元数据
Slug h-ear
版本 1.1.7
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 33
常见问题

H-ear 是什么?

H-ear.world transforms sound into an actionable, meaningful translation layer of the world around you. Describe, share and act upon environmental audio as a... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 522 次。

如何安装 H-ear?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install h-ear」即可一键安装,无需额外配置。

H-ear 是免费的吗?

是的,H-ear 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

H-ear 支持哪些平台?

H-ear 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 H-ear?

由 Paul Day(@badajoz95)开发并维护,当前版本 v1.1.7。

💬 留言讨论