← 返回 Skills 市场
375
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install indices
功能描述
Use this skill for any operation with the Indices CLI — authentication, tasks, runs, and secrets.
使用说明 (SKILL.md)
Indices CLI
Global Flags
Available on every command:
--output markdown|json(default:markdown) — usejsonfor scripting--api-base \x3Curl>(default:https://api.indices.io)--timeout \x3Cseconds>(default:30)
Auth
indices login # prompts securely for API key
indices login --api-key "\x3Ckey>" # non-interactive
indices auth-test # verify stored credentials
indices logout # remove stored API key
Tasks
Create
indices tasks create \
--display-name "Apply to jobs" \
--website "https://jobs.example.com" \
--task "Fill and submit the form"
Never set
is_fully_autonomoustotrue.
After creation, the task requires manual completion in a browser. Direct the user to https://platform.indices.io/tasks/{task_id} (substituting the actual task ID).
Flags: --display-name (required), --website (required), --task (required), --input-schema, --output-schema, --creation-params \x3Cjson-object>
JSON input (alternative to flags)
indices tasks create --body '{"display_name":"...","website":"...","task":"..."}'
indices tasks create --file ./task.json
cat task.json | indices tasks create
Rules: use at most one of --body, --file, --stdin; do not mix with argument-mode flags.
List / Get / Retry / Regenerate / Delete
indices tasks list
indices tasks list --status ready --limit 20 # statuses: not_ready | waiting_for_manual_completion | ready | failed
indices tasks get \x3Ctask-uuid>
indices tasks retry \x3Ctask-uuid>
indices tasks regenerate-api \x3Ctask-uuid>
indices tasks delete \x3Ctask-uuid> # prompts for confirmation
indices tasks delete \x3Ctask-uuid> --yes
Note: --cursor is unsupported and returns an error.
Runs
Create
indices runs create \
--task-id "\x3Ctask-uuid>" \
--arguments '{"key":"value"}' \
--secret-bindings '{"login":"\x3Csecret-uuid>"}'
Flags: --task-id (required), --arguments \x3Cjson-object>, --secret-bindings \x3Cjson-object>
JSON input follows the same rules as tasks create.
List / Get
indices runs list --task-id \x3Ctask-uuid> # --task-id required
indices runs list --task-id \x3Ctask-uuid> --limit 20
indices runs get \x3Crun-uuid>
Secrets
indices secrets create MY_SECRET --value "..." # explicit value
echo "..." | indices secrets create MY_SECRET --stdin
indices secrets create MY_SECRET # prompts securely
indices secrets list
indices secrets delete \x3Csecret-uuid> # prompts for confirmation
indices secrets delete \x3Csecret-uuid> --yes
Empty secret values are rejected.
安全使用建议
This skill is an instruction-only wrapper for the Indices CLI and will run 'indices' commands on your behalf. Before installing: (1) confirm the 'indices' CLI is installed and from a trusted source or ask the skill author to include an install spec; (2) verify which API key will be used and restrict its permissions (use a least-privilege key or a test account), because the CLI can create/delete secrets and runs; (3) be cautious about allowing autonomous invocation — if you allow it, the agent could perform destructive actions against your Indices account; (4) if you plan to use this skill, ask the publisher to update the manifest to declare the required binary and provide install instructions so the behavior is auditable.
功能分析
Type: OpenClaw Skill
Name: indices
Version: 0.0.1
The skill defines interactions with an 'indices' CLI tool, which includes powerful capabilities that could be abused. Specifically, the `--api-base <url>` flag allows specifying an arbitrary API endpoint, potentially redirecting sensitive data (like API keys or created secrets) to a malicious server if the agent is prompted to use it. Additionally, the `indices secrets create MY_SECRET --value "..."` command allows the agent to store arbitrary values as secrets, which could be exploited to exfiltrate sensitive information if the agent is instructed to retrieve and store such data. While the skill itself does not contain explicit malicious instructions or prompt injections, these documented capabilities present significant vulnerability risks if the agent is compromised by a malicious user prompt, making it suspicious.
能力评估
Purpose & Capability
The skill claims to operate the Indices CLI (login, tasks, runs, secrets) but the registry metadata lists no required binaries and provides no install spec. The instructions assume an 'indices' executable on PATH; either the manifest is incomplete or the skill will fail or behave unexpectedly.
Instruction Scope
SKILL.md instructions stay on-topic (commands for auth, tasks, runs, secrets against api.indices.io). There are no instructions to read unrelated system files, traverse arbitrary paths, or send data to endpoints outside the expected Indices API/platform URLs.
Install Mechanism
There is no install spec (instruction-only), which is low-risk. Note: because there is no installation guidance, the user or environment must already provide the 'indices' CLI; absence of that guidance is the primary concern rather than an active install risk.
Credentials
No environment variables or credentials are declared, which is consistent with an instruction-only wrapper. However, the workflow requires an API key managed by the CLI (login stores it locally). The manifest doesn't state where or how credentials are stored or recommend least-privilege keys — users should ensure the stored API key has limited scope.
Persistence & Privilege
always:false (default) and autonomous invocation is allowed (platform default). The skill will cause the agent to run the 'indices' CLI using whatever API key is present — this can perform sensitive actions (create/delete secrets, runs). This is expected for a CLI skill but worth noting before granting autonomous execution.
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 indices - 安装完成后，直接呼叫该 Skill 的名称或使用
/indices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
v0.0.1
- Initial release of the Indices CLI skill.
- Supports authentication commands: login, auth-test, logout.
- Manage tasks: create (with flags or JSON), list, get, retry, regenerate API key, delete.
- Manage runs: create (with arguments and secret bindings), list (per task), get.
- Manage secrets: create (via value, prompt, or stdin), list, delete.
- Global flags for output format, API base, and timeout supported.
元数据
常见问题
Indices 是什么？
Use this skill for any operation with the Indices CLI — authentication, tasks, runs, and secrets. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 375 次。
如何安装 Indices？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install indices」即可一键安装，无需额外配置。
Indices 是免费的吗？
是的，Indices 完全免费（开源免费），可自由下载、安装和使用。
Indices 支持哪些平台？
Indices 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 Indices？
由 itsaphel（@itsaphel）开发并维护，当前版本 v0.0.1。
推荐 Skills