skills/pointdotcom-marketing/marketing-skills/point-paid-social-template-production/Gen Agent Trust Hub
point-paid-social-template-production
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a local Python script (
scripts/preflight.py) to perform mechanical validation of PowerPoint decks and PNG images. The script uses standard libraries (zipfile,xml.etree.ElementTree,struct) to inspect file properties and structure. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external campaign briefs and user assignments, which presents a surface for indirect prompt injection.
- Ingestion points: The agent ingests campaign-specific instructions, copy, and visual references during the "Normalize the assignment" phase described in
SKILL.md. - Boundary markers: The
SKILL.mdworkflow explicitly instructs the agent to separate "unverified claims, testimonials, ratings, product facts, or legal language" from approved production assets. - Capability inventory: The skill allows the agent to execute the
scripts/preflight.pytool, write files to the local system, and generate media assets. - Sanitization: The skill mandates a manual human-review process outlined in
references/qa-preflight.mdand automated checks via the preflight script to ensure content consistency and safety.
Audit Metadata