rules-distill

Pass

Audited by Gen Agent Trust Hub on Oct 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to ingest content from multiple SKILL.md files and use an LLM to distill them into global rules. This creates an attack surface for indirect prompt injection where a malicious skill could include deceptive principles intended for system-wide adoption.
  • Ingestion points: scripts/scan-skills.sh and scripts/scan-rules.sh gather text from existing skills and rule files.
  • Boundary markers: The prompt provided to the subagent uses markdown headers (## Input) to delineate the ingested data from the analysis instructions.
  • Capability inventory: The skill performs read operations on local skill/rule directories and generates rule drafts. Any modification to rule files requires explicit, step-by-step user approval.
  • Sanitization: No automated sanitization or filtering is performed on the ingested skill content. The process relies on an LLM to identify principles and a human user to review the final drafts before they are applied.
  • [COMMAND_EXECUTION]: The skill executes local bash scripts (scripts/scan-skills.sh, scripts/scan-rules.sh) to perform inventory and metadata extraction tasks. These scripts utilize standard system utilities like find, grep, awk, and jq to process files in the user's home directory.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 14, 2026, 02:58 PM