clean-slop

Pass

Audited by Gen Agent Trust Hub on May 3, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes code diffs from Git branches, which serves as a potential vector for indirect prompt injection if the code being cleaned contains malicious instructions embedded as comments.
  • Ingestion points: Ingests untrusted code content via git diff main...HEAD as described in SKILL.md.
  • Boundary markers: No delimiters or instructions are provided to the agent to ignore instructions found within the diff data.
  • Capability inventory: The agent can perform file system modifications and execute shell commands (git).
  • Sanitization: No validation or sanitization is performed on the ingested diff content.
  • [PROMPT_INJECTION]: There is a discrepancy between the version in the frontmatter (3.0.0) and the body (v2.88). This metadata inconsistency can be deceptive or misleading regarding the skill's actual version and features.
Audit Metadata
Risk Level
SAFE
Analyzed
May 3, 2026, 11:06 AM