marketing-ideas
Pass
Audited by Gen Agent Trust Hub on Mar 6, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: No malicious patterns such as prompt injection, obfuscation, or credential theft were detected. The skill's behavior is consistent with its stated purpose of providing marketing advice.
- [NO_CODE]: The skill consists entirely of Markdown files providing textual instructions and a reference library. There are no associated Python scripts, Node.js packages, or shell commands.
- [DATA_EXPOSURE]: The skill attempts to read a local context file at
.claude/product-marketing-context.md. This is a standard pattern for maintaining project-specific context and does not involve exfiltration of sensitive system files or credentials. - [INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface as it reads
.claude/product-marketing-context.md(specified in SKILL.md). While this file could theoretically contain malicious instructions if modified by an attacker, the skill lacks any dangerous capabilities (such as network access or shell execution) that could be exploited via this vector.
Audit Metadata