prd-v01-problem-framing

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection as it is designed to ingest and process untrusted data from external sources (e.g., Reddit, G2, Capterra) using the WebSearch and WebFetch tools.
  • Ingestion points: Research data fetched from third-party websites during Step 1 and the research sequence in references/research-prompts.md.
  • Boundary markers: None identified in the prompt instructions to distinguish untrusted data from system instructions.
  • Capability inventory: The skill has the ability to write and edit local files (Write, Edit tools used for CFD-* entries and PRD.md).
  • Sanitization: No explicit sanitization or validation of the fetched external content is defined before it is processed or written to the filesystem. This is a low-risk observation consistent with the skill's primary function of information gathering.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 09:42 PM
Security Audit — agent-trust-hub — prd-v01-problem-framing