Blackbox
Security checks across static analysis, malware telemetry, and agentic risk
Overview
This is a coherent instruction-only wrapper for the Blackbox AI coding CLI, but using it means trusting a third-party CLI/API that can edit project files and send coding context to external models.
Install this only if you trust the Blackbox CLI and are comfortable sending relevant coding context to external AI providers. Use it in a scoped project directory with version control, avoid auto-approval for sensitive changes, monitor background sessions and credit usage, and keep secrets out of prompts and repositories used with the tool.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing the skill requires trusting the external Blackbox CLI package or repository.
The skill relies on installing and running a third-party global CLI or source checkout, without a pinned version in the artifact.
Blackbox CLI installed: `npm install -g @blackboxai/cli` ... `git clone https://github.com/blackboxaicode/cli.git`
Verify the npm package or GitHub repository before installing, consider pinning a version, and install in a controlled development environment.
The CLI can use the configured Blackbox account and may incur credit usage.
The integration requires a Blackbox account credential, which is expected for this service but gives the CLI access to the user's Blackbox account and paid credits.
API key from [app.blackbox.ai/dashboard](https://app.blackbox.ai/dashboard) ... run `blackbox configure` and enter your API key
Use a dedicated API key if possible, monitor credit usage, and revoke or rotate the key if no longer needed.
Blackbox may modify code in the chosen project directory, and auto-approval could reduce opportunities to review actions first.
The skill delegates coding work through terminal commands, and it documents an auto-approval flag. This is aligned with a coding-agent skill, but it can allow broad file-changing actions if used without review.
terminal(command="blackbox --prompt 'Add JWT authentication with refresh tokens to the Express API'", workdir="/path/to/project", pty=true) ... `--yolo` | Auto-approve all actions and model switches
Use a clean git working tree, review diffs after each run, and avoid `--yolo` unless you explicitly want unattended changes.
Prompts and potentially project code may be shared with external AI services during delegation.
The workflow sends task context through multiple external model providers or agents. That is the advertised feature, but users should understand the data boundary.
dispatches tasks to multiple LLMs (Claude, Codex, Gemini, Blackbox Pro) and uses a judge to select the best implementation
Do not use this on repositories containing secrets or confidential code unless the provider settings and data policies are acceptable.
Previous prompts, code context, or decisions may be reused when resuming a task.
Checkpoint and resume features imply retained task context that can influence later runs.
Blackbox CLI has built-in checkpoint support for pausing and resuming tasks ... `blackbox --resume-checkpoint 'task-abc123-2026-03-06'`
Use `/clear` or fresh sessions for unrelated work, and avoid including secrets in prompts or files exposed to the coding agent.
Multiple coding-agent processes can keep running and consuming resources or credits until completed or stopped.
The skill explicitly supports long-running background and parallel Blackbox sessions, while also documenting polling and killing sessions.
terminal(command="blackbox --prompt 'Refactor the auth module to use OAuth 2.0'", workdir="~/project", background=true, pty=true) ... Spawn multiple Blackbox instances for independent tasks
Monitor background sessions, stop ones you no longer need, and avoid running parallel agents on overlapping working directories.