← 返回 Skills 市场
354
总下载
0
收藏
1
当前安装
8
版本数
在 OpenClaw 中安装
/install recamera-intellisense
功能描述
Register and control reCamera Pro devices from an agent — onboard cameras, pick AI detection models by name, configure rule-based triggers (AI / timer / GPIO...
安全使用建议
This skill appears to do exactly what it claims: control reCamera Pro devices via a bundled Python SDK. Before installing: 1) Review and back up any existing ~/.recamera/devices.json and ensure its permissions are restricted (chmod 600). 2) Only register devices you control — tokens are long‑lived bearer credentials. 3) If you run an MCP server or any service that shares ~/.recamera/devices.json, understand that those services will have the same tokens. 4) If you need greater assurance, review the bundled Python files (they are included) before using the skill to confirm no unexpected endpoints or behaviors. 5) Note the registry metadata omission of the config path — consider asking the skill publisher to correct the package metadata so automated governance systems account for the config file.
功能分析
Type: OpenClaw Skill
Name: recamera-intellisense
Version: 2.0.8
The skill is a legitimate and well-documented stdlib-only Python SDK for controlling reCamera Pro hardware. It includes several proactive security features, such as same-origin redirect enforcement in `_http.py` to prevent bearer token leakage and robust path sanitization in `files.py` to mitigate path traversal risks. While it possesses high-risk capabilities like file deletion, storage formatting, and GPIO control, these are strictly aligned with the stated purpose of device management and include appropriate warnings and validation logic.
能力标签
能力评估
Purpose & Capability
The name/description (reCamera device management, detection, captures, GPIO, etc.) line up with the bundled Python SDK and the commands it exposes. One inconsistency: the top-level registry metadata you provided listed no required config paths, but SKILL.md metadata and the SDK both use a credential store at ~/.recamera/devices.json — so the skill will read/write that file. That config path is coherent with the skill's purpose but the registry metadata omission is a packaging inconsistency.
Instruction Scope
SKILL.md instructs the agent to run the bundled stdlib-only Python SDK and to call specific commands; the runtime instructions limit actions to device registration, HTTP/Unix-socket API calls to user-registered devices, file/capture operations, and GPIO. The SDK validates absolute paths and enforces same-origin redirects for bearer tokens. There are no instructions to read unrelated host files or to exfiltrate arbitrary data.
Install Mechanism
There is no install step (instruction-only runtime). The package includes an embedded stdlib-only Python SDK (no pip/npm downloads or external installers). That is low-risk — code runs in-process from the skill bundle rather than fetching remote code at install time.
Credentials
The skill does not request environment variables or unrelated credentials, which is appropriate. It does persist long-lived bearer tokens (format sk_...) into ~/.recamera/devices.json (stored in plaintext, auto-created with an intended chmod 600). Persisting long-lived device tokens is necessary for operation, but it is sensitive: users should ensure the file permissions, avoid committing it to VCS, and understand these tokens grant remote control of devices.
Persistence & Privilege
The skill is not force-enabled (always=false) and does not request elevated system privileges. It will create and modify ~/.recamera/devices.json and may connect to a local daemon socket (/dev/shm/rcisd.sock) if present — both are reasonable for a device-management skill. Note the SKILL.md explicitly says the config file is shared with the MCP server if installed; that sharing is an intentional design choice and increases the blast radius of leaked tokens.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install recamera-intellisense - 安装完成后，直接呼叫该 Skill 的名称或使用
/recamera-intellisense触发 - 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
v2.0.8
- Internal logic improvements made in `scripts/recamera_intellisense/rule.py`.
- No changes to public interfaces or skill behavior for users.
v2.0.7
- Selector commands now accept `name` as an alias for `device_name`, allowing direct use of records from `list_devices`.
- Clarified the required format for the `output` parameter in `start_capture` (must be an absolute on-device directory; can be omitted to use the selected storage slot).
- Improved documentation for disabling detection schedules via empty or omitted values.
- Note: requires a pending release hardware and beta software.
v2.0.6
- Adds strict same-origin redirect enforcement to prevent SSRF and token leakage on device HTTP redirects.
- Documents that HTTPS devices now validate certificates by default (secure-by-default TLS); use "allow_unsecured" only for trusted networks with self-signed certs.
- Documents that HTTP transport remains the default, but production use should switch to HTTPS.
- Documents stricter client-side path hygiene for file operations, rejecting unsafe paths before transmission.
- Notes that `~/.recamera/devices.json` is auto-created with strict (chmod 600) permissions.
- Various improvements to security and requirements documentation for clarity and safe usage.
- Note: requires and pending release hardware and beta firmware.
v2.0.5
- Minimum Python requirement raised from 3.8+ to 3.9+.
- Clarified MP4/video capture flow: for `MP4`, poll `get_capture_status` and fetch output manually; `stop_capture` only applies to in-flight video.
- Various small clarifications to requirements and capture documentation.
- Note: requires an pending release device and beta firmware.
v2.0.3
- Updated the skill description and documentation for clarity, covering reCamera Pro support, invocation, security, and command structures.
- Improved command catalogue and examples for CLI and Python usage.
- Note: requires an pending release camera and beta firmware.
v2.0.0
# reCamera Intellisense v1.0.2
Major internal reorganization and CLI rewrite.
- Reimplemented all operations as submodules under `scripts/recamera_intellisense/`, organizing features by domain (device, detection, capture, storage, GPIO, etc.).
- Added Python CLI entrypoint (`python3 -m recamera_intellisense ...`) with unified JSON I/O and command listing.
- Expanded operation set: now covers device registration, advanced detection/model/rule controls, event polling, snapshot/capture, file/record access, storage configuration, and GPIO manipulation.
- Removed legacy monolithic script files and reference docs, consolidating everything under a maintainable package layout.
- Updated documentation to reflect new workflows, CLI usage, and the relationship between MCP and CLI transports.
v1.0.1
Requires an unreleased reCamera (v2) hardware with an experimental firmware, stay tuned for the public release :-)
v1.0.0
Requires an unreleased reCamera (v2) hardware with an experimental firmware, stay tuned for the public release :-)
元数据
常见问题
reCamera Intellisense 是什么？
Register and control reCamera Pro devices from an agent — onboard cameras, pick AI detection models by name, configure rule-based triggers (AI / timer / GPIO... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 354 次。
如何安装 reCamera Intellisense？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install recamera-intellisense」即可一键安装，无需额外配置。
reCamera Intellisense 是免费的吗？
是的，reCamera Intellisense 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
reCamera Intellisense 支持哪些平台？
reCamera Intellisense 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 reCamera Intellisense？
由 iChizer0（@ichizer0）开发并维护，当前版本 v2.0.8。
推荐 Skills