slice-and-dice
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, potentially untrusted content such as technical audits, RFCs, proposals, and code architectures.
- Ingestion points: The skill ingests data from local repository files and external documents provided in the user prompt (as described in SKILL.md under 'Choosing the Right SND Mode').
- Boundary markers: The skill uses a structured sequential slicing protocol (
[SND 1/N]) and specific formatting for edits (Before→After+What Changed) which helps the user distinguish between source content and agent suggestions. However, there are no explicit instructions for the agent to ignore instructions embedded within the processed text. - Capability inventory: The skill has the capability to write approved text to target documents (Phase 4) and use interaction tools such as
ask_questionorAskUserQuestion. - Sanitization: The skill implements a robust human-in-the-loop mechanism (Phase 2), requiring explicit user confirmation (
next,apply,looks good) before any modifications are committed to the repository, which significantly mitigates the risk of an automated injection attack.
Audit Metadata