The Agent Skills Directory

[DATA_EXFILTRATION]: The skill reads data from local files such as brand-profile.json and ADS-AUDIT-REPORT.md. This is required for its stated purpose of generating creative briefs based on brand identity and audit results.
[PROMPT_INJECTION]: The skill acts as an orchestrator that ingests data from external files and user input, creating a surface for indirect prompt injection targeting downstream agents.
Ingestion points: Reads data from brand-profile.json, ADS-AUDIT-REPORT.md, ads/references/copy-frameworks.md, and skills/ads-plan/assets/ecommerce-creative.md (SKILL.md, Step 1, 2, 4, 5).
Boundary markers: Absent. The skill does not use specific delimiters or instructions to prevent instructions potentially embedded in the ingested data from being interpreted by the agent or the secondary agents it spawns.
Capability inventory: The skill writes a local file (campaign-brief.md) and spawns/tasks two secondary agents (creative-strategist and copy-writer) via the Task tool (SKILL.md, Step 5).
Sanitization: Absent. There is no evidence of validation or sanitization of the content read from the files before it is processed or passed to other agents.

ads-create