T08 · Insecure Dependencies
- Location
SKILL.md:164- Finding
Unpinned Third-Party Package Download and Execution Through npx
- Content
View full analysis
- Remediation
View remediation
Security audit
Security checks for vulnerabilities and agentic risk
This AEM API skill is mostly a plain endpoint guide, but it exposes powerful administrative and destructive actions without enough scoping, confirmation, or safe-use guidance.
Install only if you intend the agent to help with privileged AEM administration. Use least-privilege tokens, verify the target AEM host before any request, require explicit confirmation for POST or DELETE operations, avoid production until tested, and pin or review the lapsh package before running the npx commands.
SKILL.md:164Unpinned Third-Party Package Download and Execution Through npx
SKILL.md:24State-Changing AEM Request Included in Default Setup Procedure
The skill maps natural-language requests directly to a destructive endpoint, DELETE /etc/replication/agents.{runmode}/{name}, without any guardrails around parameter validation, scope restriction, or confirmation. In an agent workflow, this creates a clear pathway for tool-parameter abuse where a user or prompt injection could cause deletion of replication agents, disrupting content publishing or platform operations.
- "List all filter.tidy.2.json?" -> GET /etc/packages/{group}/{name}-{version}.zip/jcr:content/vlt:definition/filter.tidy.2.json
- "Get agents.{runmode}.-1.json details?" -> GET /etc/replication/agents.{runmode}.-1.json
- "Get agents.{runmode} details?" -> GET /etc/replication/agents.{runmode}/{name}
- "Delete a agents.{runmode}?" -> DELETE /etc/replication/agents.{runmode}/{name}
- "Create a treeactivation.html?" -> POST /libs/replication/treeactivation.html
- "Create a authorizable?" -> POST /libs/granite/security/post/authorizables
- "Create a setpassword.jsp?" -> POST /crx/explorer/ui/setpassword.jsp
The skill prominently enumerates highly sensitive AEM administrative and security-affecting endpoints, including password reset, truststore, package management, replication, auth, and system console operations, but provides no cautionary guidance, authorization checks, or safe-use constraints. In an agent setting, this can normalize dangerous actions and increase the chance the agent will assist with impactful administrative changes or destructive operations without adequate user confirmation.
npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.