outdated-deps
Pass
Audited by Gen Agent Trust Hub on Oct 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute several well-known and standard security scanners, including
npm audit,pip-audit,trivy,osv-scanner, andcargo-audit. These tools are used to detect known vulnerabilities (CVEs) within the project's dependency tree. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it is designed to ingest and analyze untrusted project data.
- Ingestion points: The workflow involves parsing external dependency manifests (e.g.,
package.json,requirements.txt,Cargo.toml) and lockfiles inSKILL.md(Step 1). - Boundary markers: There are no explicit instructions for the agent to treat manifest data as untrusted or to use specific delimiters to prevent embedded instructions from influencing agent behavior.
- Capability inventory: The skill utilizes shell command execution (to run scanners) and file system reads (Grep patterns in
references/detection-patterns.md). - Sanitization: The skill normalizes scanner output to a standard schema, but Claude-based analysis in Step 4 directly processes raw text from manifests which could contain adversarial content.
Audit Metadata