skills/megastep/codex-skills/ads-meta/Gen Agent Trust Hub

ads-meta

Pass

Audited by Gen Agent Trust Hub on Feb 26, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected in the skill files. The skill's operations are limited to reading specific local reference files and processing marketing data.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to its data ingestion design.
  • Ingestion points: The skill is designed to process external Meta Ads Manager exports and Events Manager screenshots as described in the 'Process' section of SKILL.md.
  • Boundary markers: No explicit delimiters or 'ignore embedded instruction' warnings are defined to separate ingested data from the agent's reasoning.
  • Capability inventory: The skill reads local reference documentation (e.g., ../ads/references/meta-audit.md) and generates textual reports based on its findings.
  • Sanitization: There is no indication that the skill validates or sanitizes external marketing data before incorporating it into the audit process.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 26, 2026, 01:29 AM