技能 insforge-cli

📦

insforge-cli

Name: insforge-cli
Author: insforge

内容修订 r1 严重 📁 文件系统访问🌐 网络访问🔑 环境变量⚙️ 外部命令⚡ 包含脚本

Manage InsForge Backends with the CLI

Backend administration spans many commands and safeguards. This skill guides InsForge CLI workflows for projects, databases, deployments, integrations, and diagnostics.

支持: Claude Codex Code(CC)

⚠️ 38 差

请勿自动安装此 Skill。

权威策略要求在执行任何安装操作前由操作人员审查。

使用我的智能体安装

将此请求复制给你的智能体，其中包含权威 Skill 页面和清单。

智能体请求

Review the Skillstore skill "insforge-cli" from https://skillstore.io/skills/insforge-insforge-cli.md and its manifest at https://skillstore.io/api/skills/insforge-insforge-cli/manifest. Verify the artifact. Do not auto-install. Inspect the skill and report your findings, then wait for an operator or manual installation decision.

你的智能体仍应展示其计划，并按安全策略请求必要确认。

Agent 可读资源

当 AI Agent、爬虫或脚本需要干净上下文，而不是读取完整页面时，请使用这些链接。

Markdown 详情 GET /skills/insforge-insforge-cli.md 已签名 manifest GET /api/skills/insforge-insforge-cli/manifest 已签名 lockfile GET /api/skills/insforge-insforge-cli/lockfile

测试它

正在使用“insforge-cli”。 Check whether my linked project is ready for a database migration.

预期结果:

Project status: ready. Authentication and project linkage are valid. Review the pending migration, confirm the target branch, then apply and verify it.

正在使用“insforge-cli”。 Deploy my frontend and confirm its environment settings.

预期结果:

The local build passed. Required public connection variables are configured, excluded secret files were not uploaded, and the deployment reached a healthy state.

正在使用“insforge-cli”。 Review a destructive branch reset before running it.

预期结果:

The reset restores the branch database to its original snapshot and removes later data changes. Confirm the branch name and preserve needed data first.

安全审计

严重

v3 • 2026/8/31 打开版本化报告

All 400 presented static findings were adjudicated. Confirmed risks include remote script execution, unpinned package execution, broad credential handling, and automatic transmission or retention of project context. No prompt injection or executable obfuscation was found, but nine capped static matches remain unavailable for adjudication. Static review was capped at 400/409 representative findings; omitted static matches are unconfirmed, so automatic publishing stays disabled until manual review.

已扫描文件

4,123

分析行数

复核项

已忽略误报

已确认安全问题 (13)

严重

references/compute-deploy.md:32

Pipe to shell pattern

# Requires `flyctl` on PATH (curl -L https://fly.io/install.sh | sh). NO Docker daemon needed.

The documentation directly recommends downloading a remote installation script and piping it to a shell. A compromised response or upstream host would gain immediate code execution.

严重

references/compute-deploy.md:243

Pipe to shell pattern

| `flyctl is required for source-mode deploy` | flyctl isn't installed or not on PATH | Install: `cu

The troubleshooting guidance repeats a remote script piped directly into a shell. It provides no version pin, checksum, or inspection step before execution.

高

references/login.md:86

Credential JSON file

Credentials are saved to `~/.insforge/credentials.json` with restricted file permissions (0600). The

The CLI stores bearer credentials, including full-access user API keys, in a predictable home-directory file. Mode 0600 limits access but does not protect secrets from processes running as the user.

高

references/compute-deploy.md:52

Environment file access

| `--env-file <path>` | Standard `.env` file (KEY=VALUE per line; `#` comments, blank lines, quoted

The option causes the CLI to read a local environment file that can contain deployment secrets and transmit its values to cloud infrastructure. This is legitimate but security-sensitive credential handling.

高

references/compute-deploy.md:80

Environment file access

# Env vars from a .env file (preferred for >1 secret)

The example recommends loading multiple secrets from a local environment file for cloud deployment. Accidental file selection or excessive file permissions can expose all included credentials.

高

references/compute-deploy.md:85

Environment file access

--env-file ./.env.production

This command reads the complete production environment file and sends its values during deployment. Production environment files commonly contain high-impact credentials.

高

references/local.md:65

Environment file access

`.insforge/checkout/.env` holds the generated secrets. `local start` writes

Generated database and admin secrets are persisted in a predictable local environment file. The provided gitignore reduces commit risk but does not prevent local reads or accidental disclosure.

高

references/webscraper/apify.md:49

Environment file access

- **Edge function:** `INSFORGE_BASE_URL` and `API_KEY` (the project admin key) are injected automati

The edge-function runtime automatically receives the full project admin key through its environment. Compromise of the function can therefore expose a credential with broad project access.

显示全部 13 条已确认问题

高

references/compute-deploy.md:85

Environment variant files

--env-file ./.env.production

The production environment file is explicitly consumed by the deployment command. This creates real exposure risk if the file contains unrelated secrets or is handled with weak permissions.

高

SKILL.md:14

Unpinned Automatic NPM Package Execution

The skill requires every command to use npx with automatic approval and no package version. This can download and execute a newly compromised release without review.

The line explicitly mandates npx -y for an unversioned package, which permits automatic retrieval and execution of the current registry release.

高

references/local.md:60-63

Remote Setup Script Executed by Local Workflow

The local workflow fetches a setup script from a repository and runs the generated stack. The documentation does not specify an immutable revision or integrity verification.

The reference explicitly states that the first start fetches setup.sh and runs the compose file it writes, without documenting a pinned commit or checksum.

中

SKILL.md:21 SKILL.md:332-364

Automatic Feedback Can Transmit Project Context

The skill directs agents to submit feedback automatically, including commands, errors, and attached project or organization context. Pattern-based redaction cannot guarantee removal of all user data.

The instructions require reporting before continuing and state that linked project or organization context is attached automatically. The same section acknowledges incomplete pattern-based redaction.

中

SKILL.md:20 SKILL.md:230-236

Durable Project Memory Without Per-Write Consent

The skill directs agents to recall and persist project decisions and problems as a default behavior. Sensitive project context could be stored beyond the current session.

The text explicitly requires memory lookup at task start and immediate durable storage of decisions and problems, without requiring user approval for each write.

能力复核项 (17)

这些是真实的本地能力，对此技能可能是预期行为，因此需要复核，但不会按已确认恶意行为计分。

高

references/compute-deploy.md:95

Database connection strings

--env-set DATABASE_URL=postgres://new-host \

The guidance places a database connection string directly in command arguments. Real credentials substituted here can remain in shell history and may be visible in process listings.

高

references/compute-deploy.md:101

Database connection strings

npx -y @insforge/cli compute update <id> --env '{"NODE_ENV":"production","DATABASE_URL":"..."}'

The database connection string is supplied inline as a command argument. Replacing the placeholder with a live URL can disclose credentials through shell history or process inspection.

高

references/compute-deploy.md:72

Generic API/secret keys

--env '{"OPENAI_API_KEY": "sk-..."}'

The example passes an API key in an inline command argument. A live value can be retained in shell history and exposed to local process inspection.

高

references/compute-deploy.md:96

Generic API/secret keys

--env-set API_KEY=sk-... \

The API key is supplied directly through a command-line flag. Using a real value creates avoidable exposure through history, logs, and process metadata.

高

references/login.md:67

Generic API/secret keys

npx -y @insforge/cli login --user-api-key "$INSFORGE_USER_API_KEY"

The shell expands the full-access user API key into the CLI argument vector. Environment storage avoids a literal history entry, but the expanded value may still be visible to local process inspection.

高

references/login.md:88

Generic API/secret keys

- User API key — `user_api_key` (the `uak_`, used directly as the bearer)

The full-access user API key is stored as a reusable bearer credential in the local credential file. Theft of that file permits account-level access until revocation.

高

references/login.md:99

Generic API/secret keys

npx -y @insforge/cli login --user-api-key "$INSFORGE_USER_API_KEY" --json

The login command expands a full-access API key into its argument vector. Other processes running under the same user may be able to observe that value.

高

references/webscraper/apify.md:49

Generic API/secret keys

- **Edge function:** `INSFORGE_BASE_URL` and `API_KEY` (the project admin key) are injected automati

The function receives the project admin key and uses it as a bearer credential. Any function-level code execution or logging flaw can expose broad project authority.

高

references/webscraper/apify.md:50

Generic API/secret keys

- **Fly compute:** nothing is auto-injected. Set both yourself at deploy time (`compute deploy --env

The workflow requires placing a project admin key into compute deployment configuration. This grants the deployed workload broad authority and increases impact if the workload is compromised.

高

SKILL.md:54

Generic API/secret keys

If not authenticated, run `npx -y @insforge/cli login` (opens a browser). For headless / agent / CI

The prescribed headless login expands a full-access API key into the process argument vector. The value can be exposed to local process inspection despite being sourced from an environment variable.

高

SKILL.md:226

Generic API/secret keys

- `npx -y @insforge/cli ai setup` fetches the linked project's active OpenRouter key and writes `OPE

The command retrieves an active OpenRouter key and writes it to a local environment file. This is real secret extraction and persistence with impact if the file is exposed.

高

references/login.md:86

Hidden file in home directory

Credentials are saved to `~/.insforge/credentials.json` with restricted file permissions (0600). The

A predictable hidden file stores reusable account credentials, including full-access API keys. Restrictive permissions mitigate cross-user reads but not compromise of the user account.

中

references/compute-deploy.md:85

Hidden file access

--env-file ./.env.production

The deployment command reads a production environment file likely to contain multiple secrets. This is intended behavior but gives the CLI access to sensitive local data.

中

references/local.md:65

Hidden file access

`.insforge/checkout/.env` holds the generated secrets. `local start` writes

The local workflow stores generated administrative and database secrets under a hidden project directory. Predictable storage can be read by any process with the user's filesystem privileges.

中

references/login.md:86

Hidden file access

Credentials are saved to `~/.insforge/credentials.json` with restricted file permissions (0600). The

The CLI writes and later reads reusable bearer credentials from a hidden home-directory file. This creates a real credential-theft target despite mode 0600.

中

references/payments/razorpay.md:13

Configuration library

If Razorpay is unconfigured, add Key ID and Key Secret. `config set` validates the keys and automati

The setup workflow accepts a Razorpay key secret through command-line configuration. A live secret can be retained in shell history or exposed through process metadata.

中

references/payments/stripe.md:13

Configuration library

If Stripe is unconfigured, add the environment key. `config set` validates the key and automatically

The setup workflow passes a Stripe secret key directly as a command argument. A substituted live key can leak through shell history, logs, or process inspection.

风险因素

📁 文件系统访问 (10)

references/branch/merge.md:111 references/branch/merge.md:117 references/compute-deploy.md:85 references/deployments/deploy.md:8 references/local.md:65 references/local.md:66 references/login.md:36 references/login.md:86 references/login.md:36 references/login.md:86

🌐 网络访问 (20)

🔑 环境变量 (20)

⚙️ 外部命令 (50)

⚡ 包含脚本 (2)

references/functions-deploy.md:53 references/functions-deploy.md:88

检测到的模式

Pipe to shell pattern×2

审计者: codex 查看审计历史 →

分享与引用此报告

分享版本化评估报告、中性徽章、嵌入卡片和引用。Skillstore 只报告证据，不判断该 Skill 是否安全。

打开版本化报告

复制报告链接

https://skillstore.io/skills/insforge-insforge-cli/audits/3?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown 徽章

[![Skillstore security assessment](https://skillstore.io/badges/skills/insforge-insforge-cli/security.svg)](https://skillstore.io/skills/insforge-insforge-cli?utm_source=security_passport_badge)

HTML 徽章

<a href="https://skillstore.io/skills/insforge-insforge-cli?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/insforge-insforge-cli/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

嵌入卡片

<iframe src="https://skillstore.io/embed/skills/insforge-insforge-cli.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>

学术引用 (APA · BibTeX · CFF)

APA 引用

insforge. (2026). insforge-cli security audit report (audit version 3) [Author version unspecified]. Skillstore. https://skillstore.io/skills/insforge-insforge-cli/audits/3

BibTeX 引用

@techreport{insforge-insforge-insforge-cli-2026,
  author = {insforge},
  title = {insforge-cli security audit report (audit version 3)},
  institution = {Skillstore},
  year = {2026},
  number = {3},
  url = {https://skillstore.io/skills/insforge-insforge-cli/audits/3},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "insforge-cli security audit report (audit version 3)"
version: "unspecified"
type: report
authors:
  - name: "insforge"
date-released: "2026-08-31"
url: "https://skillstore.io/skills/insforge-insforge-cli/audits/3"
identifiers:
  - type: other
    value: "skillstore:insforge-insforge-cli:audit:3"
    description: "Skillstore immutable audit report identifier"

Skillstore 评分

评分依据证据置信度: 中

架构

可维护性

内容

社区

规范符合性

你能构建什么

Provision a Backend

Create or link a project, configure services, and return verified connection details for application development.

Operate Cloud Infrastructure

Deploy applications and functions, manage compute services, configure schedules, and inspect production health.

Govern Data and Access

Plan migrations, apply row-level security, test branches, review advisor findings, and manage organization access.

试试这些提示

Inspect My Project

Check the current InsForge project. Summarize its status, linked identity, and any setup required before I continue.

Create a Database Change

Create a migration for [change]. Include grants, row-level security, integrity checks, and a safe verification plan.

Deploy an Application

Prepare and deploy [application path] to InsForge. Verify environment variables, run a local build, deploy, and report the resulting status.

Plan a Production Change

Assess [production change]. Use an isolated branch when needed, define rollback steps, request approval for destructive actions, and verify logs and advisor results.

最佳实践

Confirm the active project, organization, and branch before any write or deployment.
Keep administrative keys server-only and use protected secret handling instead of inline command arguments.
Use branches, dry runs, backups, and explicit approval before destructive or billing-impacting operations.

避免

Do not bypass CLI safeguards with undocumented backend requests.
Do not expose admin keys through public environment variables, logs, source control, or frontend bundles.
Do not auto-approve deletion, reset, transfer, role, or billing changes without clear user confirmation.

常见问题

What does this skill manage?

It manages InsForge projects, databases, deployments, functions, compute, schedules, integrations, diagnostics, billing, organizations, and related CLI workflows.

Does it write application code using the InsForge SDK?

No. Use the separate InsForge application integration skill for frontend, backend, or edge code that calls the SDK.

Does it require authentication?

Cloud operations usually require authentication. Self-hosted local operations can work without platform login after the project is linked.

Can it run destructive commands?

Yes. Some commands delete data or infrastructure, so the skill requires explicit confirmation and target verification.

How should secrets be handled?

Keep full-access keys server-only, avoid public prefixes, use protected secret stores, and never place live secrets in logs or source control.

Can it manage local InsForge instances?

Yes, when the user explicitly requests Docker-based local infrastructure. The first setup requires network access and downloads runtime components.