review-software-security
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [SAFE]: The skill functions as a defensive security guidance framework. No malicious patterns, obfuscation, hardcoded credentials, or unauthorized command executions were detected across the skill's instructions or referenced documentation.
- [INDIRECT_PROMPT_INJECTION]: The skill proactively addresses the risk of indirect prompt injection by identifying untrusted data sources and prescribing security controls.
- Ingestion points: Identified in
references/supply-chain-and-agents.mdas including retrieved documents, repository comments, issue text, web content, and tool output. - Boundary markers: The skill instructs the agent to keep external inputs outside the authority used for tool permissions and privileged actions.
- Capability inventory: Guidance includes auditing for SQL injection, shell command execution, and network fetching (SSRF) within the software being reviewed.
- Sanitization: The skill mandates using synthetic secrets for impact verification and minimal quoting of adversarial text in reports to prevent accidental execution.
- [EXTERNAL_DOWNLOADS]: The skill references multiple external resources for security best practices (OWASP, NIST, SLSA, IETF, and MITRE). These are trusted sources providing normative guidance and do not include executable scripts or untrusted dependencies.
Audit Metadata