directory-submissions
Fail
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: CRITICALDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill includes a reference to a domain (
https://fitprofessionals.net) inreferences/submission-tracker-template.csvthat is blacklisted by security scanners for malware or phishing. The workflow inSKILL.md(Step 4) instructs the agent to perform network operations usingcurlto verify the status of submissions, which would lead to the agent interacting with this malicious domain if triggered by a user request. - [INDIRECT_PROMPT_INJECTION]: The skill instructions require the agent to ingest content from external files such as
.agents/product-marketing.mdand.claude/product-marketing.mdto establish context. - Ingestion points: Relative file paths
.agents/product-marketing.md,.claude/product-marketing.md, andproduct-marketing-context.mdspecified inSKILL.md. - Boundary markers: The instructions do not define clear delimiters or "ignore embedded instructions" guards for the data read from these files.
- Capability inventory: The skill includes the ability to perform shell commands (
curl) to verify network resources as described in Step 4 of the workflow. - Sanitization: There is no evidence of sanitization or validation logic for the content ingested from the product marketing context files.
Recommendations
- CRITICAL: 2 file(s) identified as malware by FileRep - DO NOT USE
- Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata