pmg-qa-audit
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes external PowerPoint (.pptx) files, which could theoretically contain malicious text or metadata designed to influence the agent's logic during the audit process.
- Ingestion points: High-complexity data is ingested from presentation files via the
pmg validateandpmg inspectcommands (SKILL.md). - Boundary markers: The instructions do not prescribe specific delimiters or safety warnings for the model to ignore natural language content found within the slides during analysis.
- Capability inventory: The skill has the ability to read from the file system and execute shell commands using the
pmgtoolset. - Sanitization: There is no mention of content filtering or sanitization for strings extracted from the presentations.
- [SAFE]: The core functionality of the skill is centered on read-only diagnostics and immutable certification of artifacts. All tools used (e.g.,
pmg.validation,pmg.diagnostics) are relevant to the stated purpose of QA auditing and do not exhibit suspicious behaviors.
Audit Metadata