social-media-post
Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows standard prompt engineering practices for complex task orchestration. It instructs the agent to read local configuration files (e.g.,
meta.json,global-constraints.json) to define its behavior, which is a benign use of the agent's file system capabilities. - [DATA_EXPOSURE]: The skill accesses project-specific files such as
brand-voice.mdandmarketing-triggers.json. These references are scoped to the skill's own directory or acore/subdirectory and do not target sensitive system paths, credentials, or user-private data. - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes user-provided
$ARGUMENTSand the contents ofbrand-voice.mdwithout explicit boundary markers or delimiters. However, the risk is negligible as the skill possesses no dangerous capabilities such as network access, file writing, or command execution; its output is restricted to generating social media post text.
Audit Metadata