Skills agent-tools

📦

agent-tools

Name: agent-tools
Author: inferen-sh

Content revision r1 Critical 🌐 Network access🔑 Env variables📁 Filesystem access⚙️ External commands

Run AI Apps with inference.sh

Teams need one way to run many AI models without managing GPUs. This skill guides Claude, Codex, and Claude Code through inference.sh CLI workflows.

Supports: Claude Codex Code(CC)

⚠️ 38 Poor

Do not auto-install this skill.

The canonical policy requires operator review before any installation action.

Install with my Agent

Copy this request to your Agent. It includes the canonical Skill page and manifest.

Agent request

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

Your Agent should still show its plan and request any confirmation required by the security policy.

Agent-readable resources

Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.

Markdown detail GET /skills/inferen-sh-agent-tools.md Signed manifest GET /api/skills/inferen-sh-agent-tools/manifest Signed lockfile GET /api/skills/inferen-sh-agent-tools/lockfile

Test it

Using "agent-tools". Search for a fast video generation app.

Expected outcome:

A shortlist of relevant inference.sh video apps, speed notes, and a recommended app to inspect next.

Using "agent-tools". Run an image generation task from a prepared input file.

Expected outcome:

A concise task report with the app name, task ID, status, and where to find the generated image result.

Using "agent-tools". Compare search providers available through inference.sh.

Expected outcome:

A practical comparison of Tavily, Exa, and related apps, including when each option is useful.

Security Audit

Critical

v4 • 7/5/2026 Open versioned report

The audit confirms critical remote installer execution through repeated curl-to-shell instructions, plus high-risk local file upload, hidden-directory install, and broad CLI permission concerns. Many documentation links, Markdown backticks, placeholder API key examples, and troubleshooting references are false positives. No prompt injection text was found in the reviewed files.

Files scanned

596

Lines analyzed

Review items

False positives ignored

Confirmed security concerns (7)

Critical

references/authentication.md:6

Pipe to shell pattern

curl -fsSL https://cli.inference.sh | sh

curl piped directly to sh executes remote installer content before inspection. A compromised endpoint or response could run arbitrary shell code.

Critical

references/authentication.md:44

Pipe to shell pattern

curl -fsSL https://cli.inference.sh | sh

The reinstall instructions repeat the same curl pipe to shell pattern. This remains arbitrary remote shell execution if the installer source is compromised.

Critical

references/cli-reference.md:6

Pipe to shell pattern

curl -fsSL https://cli.inference.sh | sh

The CLI reference installs by piping a remote curl response into sh. That pattern can execute arbitrary code from the remote installer source.

Critical

SKILL.md:16

Pipe to shell pattern

curl -fsSL https://cli.inference.sh | sh

The primary install instruction pipes a remote installer directly into sh. This is a high-risk remote code execution pattern if the endpoint or response is compromised.

High

SKILL.md:4 references/cli-reference.md:56-66

Overbroad infsh Command Permission

The skill grants Bash(infsh *) and documents commands for deployment, source pulling, app execution, and task handling. This broad wildcard can allow high-impact actions through one CLI namespace.

The allowed-tools line explicitly permits all infsh subcommands, and the CLI reference lists deployment and source-pull operations. The risk is contextual rather than a single syntax pattern.

High

SKILL.md:56-72 references/running-apps.md:21-37

Local File Uploads to Cloud Apps

The skill says local paths are automatically uploaded when used as app inputs. An agent could send unintended local image, audio, video, or media files to inference.sh apps.

Both cited sections explicitly describe automatic local file upload behavior and provide commands with local paths. This supports a strong data-disclosure risk finding.

High

SKILL.md:49-50 SKILL.md:70-71 SKILL.md:95-96

Automated Social Posting Capability

The examples include posting to Twitter/X and list actions such as post, direct message, follow, like, and retweet. Authorized misuse could publish or amplify content from a connected account.

The commands and capability table explicitly mention X posting and engagement actions. This is a business-logic abuse risk static pattern matching did not fully capture.

Capability review items (28)

These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.

High

references/cli-reference.md:84

Hidden file in home directory

infsh completion fish > ~/.config/fish/completions/infsh.fish

The command writes generated completion content into ~/.config/fish/completions. This is a user-visible setup action, but it modifies shell configuration in a hidden home path.

High

references/running-apps.md:41

Path traversal sequence

- Relative paths: `./image.png`, `../data/video.mp4`

The documentation says parent-directory paths such as ../data/video.mp4 are supported for local uploads. If an agent supplies such paths, files outside the working directory can be uploaded to the cloud.

High

SKILL.md:23-32

Ruby/shell backtick execution

> ```bash

The manual install block contains shell commands for curl, checksum verification, archive extraction, and moving a binary. These are real local and network commands users may run.

High

SKILL.md:60-72

Ruby/shell backtick execution

```bash

The command block runs infsh apps with local image, audio, and media paths. Those commands can upload local files to remote inference services.

High

SKILL.md:26

Shell command substitution

> curl -LO $(curl -fsSL https://dist.inference.sh/cli/manifest.json | grep -o '"url":"[^"]*"' | grep

The manual install command uses command substitution around a network curl pipeline to choose a download URL. A malicious or unexpected manifest response could influence the downloaded artifact.

High

SKILL.md:23-32

Template literal with command substitution

> ```bash

The manual install block contains the command substitution on line 26 inside a shell command sequence. This is a real dynamic download construction, not only Markdown formatting.

High

SKILL.md:31

Hidden file in home directory

> mv inferencesh-cli-* ~/.local/bin/inferencesh

The manual install command moves the downloaded binary into ~/.local/bin. This modifies a hidden home directory that is commonly on PATH, affecting future command execution.

Medium

references/cli-reference.md:84

Hidden file access

infsh completion fish > ~/.config/fish/completions/infsh.fish

The fish completion command targets a hidden configuration directory under the user home. It is legitimate CLI setup, but it changes shell behavior through local filesystem writes.

Medium

SKILL.md:15-18

Ruby/shell backtick execution

```bash

The fenced block instructs command execution for installer setup and login. The dangerous remote-shell aspect is captured separately, but external command execution is intended here.

Medium

SKILL.md:36-54

Ruby/shell backtick execution

```bash

The quick examples run infsh app commands against cloud AI services and X posting. This is intended external CLI execution with network side effects.

Medium

SKILL.md:78-79

Ruby/shell backtick execution

| List all apps | `infsh app list` |

The table directs use of infsh app list. It is a normal CLI command, but the skill intentionally delegates app discovery to an external command.

Medium

SKILL.md:79-80

Ruby/shell backtick execution

| Search apps | `infsh app list --search "flux"` |

The table directs use of infsh app list with a search query. This is intended external CLI execution against the inference.sh service.

Medium

SKILL.md:80-81

Ruby/shell backtick execution

| Filter by category | `infsh app list --category image` |

The table directs category filtering through the infsh CLI. It is legitimate documentation, but still an external command path.

Medium

SKILL.md:81-82

Ruby/shell backtick execution

| Get app details | `infsh app get google/veo-3-1-fast` |

The table directs fetching app details through infsh app get. This invokes an external CLI to retrieve service metadata.

Medium

SKILL.md:82-83

Ruby/shell backtick execution

| Generate sample input | `infsh app sample google/veo-3-1-fast --save input.json` |

The command generates sample input and saves it to input.json. This combines external CLI execution with local file creation.

Medium

SKILL.md:83-84

Ruby/shell backtick execution

| Run app | `infsh app run google/veo-3-1-fast --input input.json` |

The run command executes a cloud app through infsh with a local input file. This can spend credits and send input data to the remote service.

Medium

SKILL.md:84-85

Ruby/shell backtick execution

| Run without waiting | `infsh app run <app> --input input.json --no-wait` |

The no-wait run command submits a cloud task and returns before completion. It is intended external execution with remote side effects.

Medium

SKILL.md:85-101

Ruby/shell backtick execution

| Check task status | `infsh task get <task-id>` |

The task lookup command invokes infsh against a task identifier. It is ordinary CLI usage, but confirms the skill performs external command-based task management.

Medium

SKILL.md:101-127

Ruby/shell backtick execution

```bash

The related skills block uses npx skills add commands. These commands can fetch and install additional skill packages from a network source.

Medium

SKILL.md:31

Hidden file access

> mv inferencesh-cli-* ~/.local/bin/inferencesh

The install step writes into ~/.local/bin under the user home. That is legitimate setup, but it is still hidden-directory filesystem modification.

Low

references/authentication.md:6

Hardcoded URL

curl -fsSL https://cli.inference.sh | sh

The URL is used in a curl command whose output is piped to sh. This downloads executable installer content from the network during setup.

Low

references/authentication.md:44

Hardcoded URL

curl -fsSL https://cli.inference.sh | sh

The reinstall command again downloads the remote installer from cli.inference.sh and pipes it to a shell. This is a real network execution dependency.

Low

references/cli-reference.md:6

Hardcoded URL

curl -fsSL https://cli.inference.sh | sh

The hardcoded URL is part of the CLI installation command and retrieves executable installer content. It is an intended but real network trust boundary.

Low

SKILL.md:11

Hardcoded URL

![[inference.sh](https://inference.sh)](https://cloud.inference.sh/app/files/u/4mg21r6ta37mpaz6ktzwt

The Markdown image points to a cloud.inference.sh asset, which can trigger a remote fetch when rendered. This is a low-severity network privacy dependency rather than code execution.

Low

SKILL.md:16

Hardcoded URL

curl -fsSL https://cli.inference.sh | sh

The URL is used in the curl pipe to shell install command. It downloads remote executable installer content during setup.

Low

SKILL.md:20

Hardcoded URL

> **What does the installer do?** The [install script](https://cli.inference.sh) detects your OS and

The installer explanation links directly to the remote install script source. It is part of the same network trust boundary used for setup.

Low

SKILL.md:25

Hardcoded URL

> curl -LO https://dist.inference.sh/cli/checksums.txt

The manual install path downloads checksums from dist.inference.sh. This is a legitimate verification step, but it still relies on a hardcoded external endpoint.

Low

SKILL.md:26

Hardcoded URL

> curl -LO $(curl -fsSL https://dist.inference.sh/cli/manifest.json | grep -o '"url":"[^"]*"' | grep

The command downloads a remote manifest and uses it to select the CLI artifact. This is a real external network dependency in the installer path.

Risk Factors

🌐 Network access (17)

references/authentication.md:6 references/authentication.md:44 references/cli-reference.md:6 references/running-apps.md:152 SKILL.md:7 SKILL.md:11 SKILL.md:16 SKILL.md:20 SKILL.md:25 SKILL.md:26 SKILL.md:138 SKILL.md:139 SKILL.md:140 SKILL.md:141 SKILL.md:142 SKILL.md:143 SKILL.md:144

🔑 Env variables (3)

references/authentication.md:30 references/authentication.md:53 references/cli-reference.md:72

📁 Filesystem access (5)

references/cli-reference.md:84 references/cli-reference.md:84 references/running-apps.md:41 SKILL.md:31 SKILL.md:31

⚙️ External commands (20)

Detected Patterns

Pipe to shell pattern×4

Audited by: codex View Audit History →

Share & cite this report

Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.

Open versioned report

Copy report link

https://skillstore.io/skills/inferen-sh-agent-tools/audits/4?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown badge

[![Skillstore security assessment](https://skillstore.io/badges/skills/inferen-sh-agent-tools/security.svg)](https://skillstore.io/skills/inferen-sh-agent-tools?utm_source=security_passport_badge)

HTML badge

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

Embed card

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

Academic citations (APA · BibTeX · CFF)

APA citation

inferen-sh. (2026). agent-tools security audit report (audit version 4) [Author version unspecified]. Skillstore. https://skillstore.io/skills/inferen-sh-agent-tools/audits/4

BibTeX citation

@techreport{inferen-sh-inferen-sh-agent-tools-2026,
  author = {inferen-sh},
  title = {agent-tools security audit report (audit version 4)},
  institution = {Skillstore},
  year = {2026},
  number = {4},
  url = {https://skillstore.io/skills/inferen-sh-agent-tools/audits/4},
  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: "agent-tools security audit report (audit version 4)"
version: "unspecified"
type: report
authors:
  - name: "inferen-sh"
date-released: "2026-07-05"
url: "https://skillstore.io/skills/inferen-sh-agent-tools/audits/4"
identifiers:
  - type: other
    value: "skillstore:inferen-sh-agent-tools:audit:4"
    description: "Skillstore immutable audit report identifier"

Compare variants

20 installable variants

Each author remains a separate installable skill. The recommended variant is ranked by Skillstore evidence.

Why this variant is first

Higher Skillstore usage

Author / source Skillstore Score Evidence Confidence Skillstore usage Updated

tul-sh Recommended

tul-sh-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 26

Updated

2026-08-21

toolshell

toolshell-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 23

Updated

2026-08-21

inferen-sh Current

inferen-sh-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 14

Updated

2026-08-21

tool-belt

tool-belt-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 12

Updated

2026-08-21

inferencesh

inferencesh-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 11

Updated

2026-08-21

inference-sh-skills

inference-sh-skills-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 10

Updated

2026-08-21

inference-sh-7

inference-sh-7-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 10

Updated

2026-08-21

inf-sh

inf-sh-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 9

Updated

2026-08-21

inference-skills

inference-skills-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 9

Updated

2026-08-21

skillssh

skillssh-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 8

Updated

2026-08-21

inference-sh-9

inference-sh-9-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 8

Updated

2026-08-21

inference-sh-3

inference-sh-3-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 8

Updated

2026-08-21

qu-skills

qu-skills-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 7

Updated

2026-08-21

infsh-skills

infsh-skills-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 7

Updated

2026-08-21

inference-sh-6

inference-sh-6-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 7

Updated

2026-08-21

inference-sh-0

inference-sh-0-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 7

Updated

2026-08-21

inference-shell

inference-shell-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 6

Updated

2026-08-21

inference-sh-8

inference-sh-8-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 5

Updated

2026-08-21

101-skills

101-skills-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 5

Updated

2026-08-21

halt-catch-fire

halt-catch-fire-agent-tools

Skillstore Score 38

Evidence Confidence Medium

Skillstore usage 4

Updated

2026-08-21

Skillstore Score

Why this score Evidence Confidence: Medium

Architecture

Maintainability

Content

Community

Spec Compliance

What You Can Build

Prototype multimodal assets

Search image, video, audio, and 3D apps, then run selected models for rapid creative experiments.

Run model experiments

Inspect app schemas, prepare sample inputs, execute cloud jobs, and compare results across providers.

Operate task workflows

Submit long-running jobs, monitor task IDs, and save outputs for handoff or automation.

Try These Prompts

Find an app

Use inference.sh to search for image generation apps. Show the best options and note which app you recommend.

Prepare a run

Inspect falai/flux-dev-lora with infsh and generate a sample input file. Explain the required fields before I run it.

Run and monitor

Run google/veo-3-1-fast with my prepared input file, do not upload unrelated local files, and report status until completion.

Plan a workflow

Design an inference.sh workflow that creates an image, turns it into a short video, and records task IDs and costs to check.

Best Practices

Confirm the app name, account, cost, and output destination before running cloud jobs.
Use sample inputs and inspect schemas before long video, audio, or social posting tasks.
Use explicit local file paths and avoid parent directories unless the user requested the exact file.

Avoid

Do not pipe remote installers to a shell without inspection and verification.
Do not pass broad local directories or parent paths as model inputs.
Do not let an agent post to X or spend credits without explicit approval.

Frequently Asked Questions

Does this skill require an inference.sh account?

Yes. You need an inference.sh account, CLI authentication, network access, and available credits for paid apps.

Can it work with Claude, Codex, and Claude Code?

Yes. The report lists support for Claude, Codex, and Claude Code.

Will it upload my local files?

It can upload local files when a supported app input uses a local path. Confirm each file before running.

Can it generate images and videos?

Yes. It documents image and video apps such as FLUX, Veo, Seedance, and related providers.

Does it install the CLI automatically?

It provides installation commands, but installation should be reviewed and approved by the user first.

Can it monitor long-running tasks?

Yes. It documents task ID lookup, status checks, and saving results when the CLI supports it.

Developer Details

Author

inferen-sh

License

MIT

Skillstore revision

Version notice

The author did not declare a version.

Repository

https://github.com/inferen-sh/skills/tree/main/tools/agent-tools/

Ref

30c73eac2afe762f6aa9c4553158769369d47351

Maintenance freshness

7/18/2026

Usage

9 downloads · 190 views

File structure

📁 references/

📄 app-discovery.md

📄 authentication.md

📄 cli-reference.md

📄 running-apps.md

📄 SKILL.md

More from inferen-sh

View all

📦

elevenlabs-tts

Create ElevenLabs Voiceovers

Producing natural narration often requires manual setup across voice, model, and format options. This skill helps Claude, Codex, and Claude Code run ElevenLabs TTS through the inference.sh CLI.

Communication View

📦

qwen-image-2-pro

Generate Professional Images with Qwen Image Pro

Text-heavy image prompts often produce weak layouts and unreadable lettering. This skill guides Qwen-Image-2.0-Pro generation and editing through inference.sh.

Design & Creative View

📦

nano-banana-2

Generate and Edit Images with Nano Banana 2

Creating polished images often requires separate tools for generation, editing, and current information. This skill provides one inference.sh workflow for all three tasks.

Design & Creative View

📦

ai-video-generation

Create AI Videos with Inference.sh

Teams need fast access to many AI video models without learning every provider API. This skill gives Claude, Codex, and Claude Code concise inference.sh workflows.

Design & Creative View

📦

tools-ui

Build Tool Lifecycle Interfaces for React

Agent applications need clear feedback for tool calls, approvals, and results. This skill provides React patterns for presenting each lifecycle state consistently.

Coding & Development View

📦

chat-ui

Build React Chat Interfaces

Teams need chat interfaces that feel consistent with modern React applications. This skill provides reusable chat component patterns for React, Next.js, and shadcn-style projects.

Coding & Development View

View all

📦

writingmate-mcp

Compare AI Models and Generate Media

by writingmate

Choosing and testing AI models across separate services takes time. Writingmate connects an MCP host to model discovery, text comparison, image generation, and video generation.

Productivity View

📦

ponyflash

Generate and Edit AI Media with PonyFlash

by ponyflash

AI media work often spans separate generation, editing, and subtitle tools. PonyFlash combines SDK guidance with FFmpeg workflows for images, videos, speech, music, and exports.

Design & Creative View

📦

runcomfy-cli

Run RunComfy Models from the CLI

by runcomfy-com

Teams need a repeatable way to run RunComfy media models from scripts. This skill guides CLI install, authentication, model runs, polling, and downloads.

Design & Creative View

📦

generating-storyboards

Create Consistent Video Storyboards

by supercmohq

Video concepts can lose continuity when clips are planned separately. This skill creates sequential storyboard sheets that preserve characters, products, settings, and visual states.

Design & Creative View

📦

generating-ai-actors

Create Reusable Photorealistic AI Actors

by supercmohq

Campaign teams need consistent fictional talent without supplying a person photograph. This skill creates one detailed adult identity that can guide later image and video generation.

Design & Creative View

📦

imagegen

Create and Edit Project Images

by openai

Producing consistent raster assets requires precise prompts, careful edits, and reliable output handling. This skill structures requests, creates or edits visuals, validates results, and saves project-ready files.

Design & Creative View

agent-tools

Install with my Agent

Agent-readable resources

Test it

Security Audit

Confirmed security concerns (7)

Risk Factors

Detected Patterns

Copy report link

Markdown badge

HTML badge

Embed card

APA citation

BibTeX citation

CITATION.cff

Compare variants

Skillstore Score

What You Can Build

Prototype multimodal assets

Run model experiments

Operate task workflows

Try These Prompts

Best Practices

Avoid

Frequently Asked Questions

Developer Details

More from inferen-sh

elevenlabs-tts

qwen-image-2-pro

nano-banana-2

ai-video-generation

tools-ui

chat-ui

Related Skills

writingmate-mcp

ponyflash

runcomfy-cli

generating-storyboards

generating-ai-actors

imagegen