skill-crafting

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides shell command templates and a pre-written bash script intended for structural validation. These commands utilize common utilities like grep, sed, jq, and wc to analyze local markdown and YAML files for quality assurance.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it is designed to ingest and iterate upon user-provided skill definitions.
  • Ingestion points: Processes new or existing files within the local skills/ directory, specifically SKILL.md and references/guide.md.
  • Boundary markers: Absent. The skill does not specify the use of delimiters or 'ignore' markers to separate developer instructions from the content of the target skills.
  • Capability inventory: Capability to execute local shell commands for file analysis and use interactive tools like AskUserQuestion to influence agent logic.
  • Sanitization: Absent. The framework treats skill logic and instructional content as plain text, relying on manual review rather than automated sanitization.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 09:02 PM