requirement-forge

Pass

Audited by Gen Agent Trust Hub on May 26, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads and synthesizes untrusted external materials (PRDs, feature lists, Jira exports, etc.) in Step 2 of its workflow. Malicious instructions embedded in these source materials could attempt to hijack the agent's logic during synthesis or classification.
  • Ingestion points: Step 2 Intake (SKILL.md) reads user-provided files and material from the repository.
  • Boundary markers: Absent. The skill does not use delimiters or provide instructions to the model to ignore potential commands within the source text.
  • Capability inventory: The skill has file-writing capabilities, specifically creating and updating documents in the .lattice/requirements/ directory across multiple steps.
  • Sanitization: Absent. There is no mention of sanitizing or validating the content extracted from external materials before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
May 26, 2026, 11:04 PM
Security Audit — agent-trust-hub — requirement-forge