diffity-resolve
Pass
Audited by Gen Agent Trust Hub on May 11, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads and interprets review comments from an external source, creating an attack surface for indirect prompt injection.\n
- Ingestion points: The skill parses JSON output from
diffity agent list --status open --jsonto read comment bodies.\n - Boundary markers: No delimiters or specific ignore-instructions are used when processing the comment content.\n
- Capability inventory: The agent can use the Edit tool to modify source files and execute CLI commands via
diffity.\n - Sanitization: No sanitization or validation of the comment text is performed before the agent interprets the intent and applies changes.\n- [EXTERNAL_DOWNLOADS]: The skill checks for and provides instructions to install the
diffityCLI tool globally usingnpm install -g diffity. This package is central to the skill's functionality and is intended for use within the Nilbuild ecosystem.
Audit Metadata