synthesis-content-framing
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided technical articles to apply editorial quality gates. This ingestion of untrusted data creates a potential surface for indirect prompt injection.
- Ingestion points: Processes articles, blog posts, and technical thought leadership content as described in
SKILL.mdandreferences/quality-gates.md. - Boundary markers: The instructions do not specify the use of delimiters or clear directives for the agent to ignore instructions contained within the analyzed content.
- Capability inventory: The agent is instructed to perform network operations (
curl) and build commands (npm run build) in certain workflows. - Sanitization: No sanitization or input validation mechanisms are defined for the processed text.
- [COMMAND_EXECUTION]: The skill contains specific instructions for the agent to execute shell commands to verify deployment status.
- Evidence:
references/quality-gates.mddirects the agent to executecurl -o /dev/null -w "%{http_code}" "URL"to check for 404 responses and mentions usingnpm run buildduring content deletion verification.
Audit Metadata