paid-ads-integrations
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to load and adopt instructions from an external relative file path (
../../paid-ads/paid-ads-integrations/SKILL.md) as its "active workflow." - Ingestion points: The agent is directed to read the content of the target markdown file located at a relative path from the current directory.
- Boundary markers: Absent. The skill provides no delimiters or instructions to treat the external file's content as data rather than instructions; instead, it explicitly commands the agent to follow it as the primary workflow.
- Capability inventory: The wrapper file itself does not define any tools, network operations, or subprocess calls.
- Sanitization: Absent. There is no validation or filtering of the target file's content before it is adopted as the agent's active instructions.
- [DYNAMIC_EXECUTION]: The skill uses a redirection pattern to switch the agent's operational logic to a different file at runtime. Although the path is hardcoded and points to a relative local directory (
<plugin-root>/paid-ads/paid-ads-integrations/), this mechanism allows the agent's behavior to be entirely defined by an external resource rather than the skill's own file.
Audit Metadata