judgment-day

Pass

Audited by Gen Agent Trust Hub on Aug 24, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its automated processing of untrusted external data to drive code-writing agents.
  • Ingestion points: The skill explicitly reads untrusted content including diffs, file targets, PR descriptions, and Jira tickets as defined in the SKILL.md input sequence.
  • Boundary markers: The templates in references/prompts-and-formats.md for the Judge and Fix Agent prompts lack explicit boundary markers, delimiters, or negative constraints to prevent the LLM from following instructions embedded within the target files or descriptions.
  • Capability inventory: The skill possesses the capability to modify the local filesystem by delegating tasks to a Fix Agent (described in SKILL.md execution steps) that is instructed to apply changes to the codebase.
  • Sanitization: There is no evidence of sanitization, escaping, or validation of external content before it is interpolated into the prompts for the judge or fix agents.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 24, 2026, 11:18 AM
Security Audit — agent-trust-hub — judgment-day