http-header-security-audit
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill facilitates auditing HTTP headers, which involves ingesting data from untrusted external servers.
- Ingestion points: Data retrieved from external URLs and specific HTTP headers during the audit process (SKILL.md).
- Boundary markers: Absent. There are no explicit instructions for using delimiters to isolate untrusted header data from the agent's logic.
- Capability inventory: The skill frontmatter permits the use of the Bash and Write tools, which provide capabilities for network requests and file system modifications (SKILL.md).
- Sanitization: Absent. No sanitization or filtering of the external header content is described in the instructions.
Audit Metadata