ads-photoshoot

Pass

Audited by Gen Agent Trust Hub on May 8, 2026

Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [PROMPT_INJECTION]: Indirect prompt injection surface identified in the prompt construction logic. The skill interpolates user-supplied product descriptions, external URL content, and brand profile data directly into templates for image generation.
  • Ingestion points: User input (product description), external product URL (Step 1), and local brand configuration file brand-profile.json (Step 2).
  • Boundary markers: Absent. The skill lacks delimiters or instructions to ignore potential malicious content embedded within the ingested data.
  • Capability inventory: The skill has the capability to write files to the local filesystem (./product-photos/) and execute generation commands via the banana-claude tool.
  • Sanitization: No sanitization or validation of input data is performed before prompt interpolation.
  • [DATA_EXFILTRATION]: Access to sensitive file paths in the user's home directory.
  • Evidence: The skill reads from ~/.banana/costs.json to report usage spend to the user.
  • Evidence: The skill references documentation and specification files located in ~/.claude/skills/ads/references/.
Audit Metadata
Risk Level
SAFE
Analyzed
May 8, 2026, 03:43 PM
Security Audit — agent-trust-hub — ads-photoshoot