npm AI-agent policy reports, with evidence.

The package has benign runtime book-writing code, but its install hook unconditionally writes package instructions into a project-level AI-agent skills directory. Under the install contro...

The runtime library is benign local code analysis, but the lifecycle hook unconditionally plants agent instructions into a project-level .opencode skill surface. Under the install control...

Source inspection confirms automatic lifecycle installation into .opencode/skills, which falls under unconsented AI-agent control-surface mutation. The absence of exfiltration or classic...

Although the payload is a notification feature and no exfiltration is present, the package uses an npm postinstall hook to modify ~/.claude/settings.json and register Claude Code executio...

Although the content is product-aligned and there is no exfiltration or remote code execution, lifecycle-triggered writes into ~/.claude/skills are an unconsented mutation of a foreign AI...

The package's core behavior is an install-time rewrite of Claude Code's global hooks/settings, which matches the blockable AI-agent control-hijack policy even though no classic malware or...

Although the payload is product-aligned and lacks network exfiltration, it is lifecycle-triggered and unconditionally rewrites a broad foreign AI-agent control surface under ~/.claude. Un...

Static source inspection confirms install-time code plants repository-wide AGENTS.md instructions, which the supplied policy treats as blockable AI-agent control hijack when lifecycle-tri...

The concrete issue is not classic malware but lifecycle-triggered mutation of Claude Code control files in the consuming project, which the policy treats as blockable AI-agent control hij...

Static inspection confirms automatic lifecycle mutation of Claude Code .claude agents/commands, which is a foreign/broad AI-agent control surface. Even though no exfiltration, remote code...

Static inspection confirms an unconsented npm lifecycle write into ~/.claude/skills, a foreign/broad AI-agent control surface covered by the block policy. The planted content appears beni...

Static source inspection confirms unconsented lifecycle mutation of ~/.claude/skills, a broad/foreign AI-agent control surface covered by the block policy. The planted content appears saf...

Under the install-control-surface policy, unconsented lifecycle writes into ~/.claude/skills are blockable AI-agent control hijack even when the content appears product-aligned. Other net...

The package appears to be a real OpenCode/NIM booster, but its postinstall performs unconsented registration into OpenCode agent configs and project config. Under the install-control-surf...

Although the plugin functionality and Webshare endpoints are product-aligned, npm lifecycle execution unconditionally registers the package into OpenCode user/home configs and drops proje...

Static source inspection confirms unconsented lifecycle-time mutation of a broad OpenCode AI-agent control surface, which is blockable under the install-control-surface policy. The proxy...

Static inspection confirms install-time execution of setup and package documentation states that setup creates a supervised daemon plus agent plugins. Under the provided policy, unconsent...

Although the functionality is package-aligned and documented, it lifecycle-mutates a broad Claude Code control surface in the user's home config without an interactive opt-in. Under the i...

Although product-aligned and documented, the package uses an npm lifecycle hook to mutate Claude Code user settings and install an agent hook without an explicit install command. Under th...

Static inspection confirms lifecycle-triggered MCP configuration mutation plus persistent service/global install behavior, which matches the firewall policy for unconsented AI-agent contr...

Static inspection confirms unconsented postinstall writes into a foreign/broad AI-agent control surface (.claude/skills, .claude/agents, .claude/commands), which is blockable under the in...