registry / @nextsparkjs/ai-workflow / 0.1.0-beta.186

@nextsparkjs/ai-workflow@0.1.0-beta.186

AI workflow templates for NextSpark - Claude Code agents, commands, skills, and multi-editor support

AI Security Review

scanned 1mo ago · by lpm-firewall-ai

Installation can automatically replace a broad set of Claude Code control files in an existing consumer project. The guard is vendor-specific but does not obtain consent for each install or update.

Static reason

One or more suspicious static signals were detected.

Trigger

npm installation or update in a NextSpark project that already contains .claude.

Impact

Consumer AI-agent instructions, commands, skills, and workflows can be replaced without an explicit setup command.

Mechanism

Postinstall spawns setup and overwrites packaged Claude workflow files.

Attack narrative

The package registers an automatic postinstall hook. In a project with NextSpark dependencies and an existing .claude directory, it runs its setup script with that directory as the working directory. Setup copies package-controlled content over matching consumer agent, command, skill, template, workflow, and documentation files. This is an unconsented install-time mutation of a broad third-party AI-agent control surface.

AI rationale

The lifecycle hook performs automatic broad mutation of the consumer's .claude control surface. Its guards and lack of network behavior reduce scope but do not establish user consent for the mutation.

Evidence

package.jsonscripts/postinstall.mjsscripts/setup.mjs.claude/agents.claude/commands.claude/skills.claude/templates.claude/workflows.claude/_docs

@nextsparkjs/ai-workflow@0.1.0-beta.186

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Affected versions and remediation

Source & flagged code