normalize-skill-for-claude

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data by reading and auditing external skill files provided via the $target argument. A malicious skill file could contain embedded instructions designed to override the normalization logic or influence the agent's behavior during the audit and rewrite phases.
  • Ingestion points: SKILL.md resolves and reads a target skill directory or file in Step 1 and Step 3.
  • Boundary markers: The instructions do not define explicit delimiters or "ignore embedded instructions" warnings for the content read from the target skill.
  • Capability inventory: The skill utilizes Bash, Read, Write, Edit, Glob, and Grep tools to perform its tasks.
  • Sanitization: No explicit sanitization or filtering of the target skill content is mentioned before it is processed by the agent.
  • [COMMAND_EXECUTION]: The skill makes legitimate use of the Bash tool to confirm the existence of target files, inventory directories, and perform validation checks on the rewritten artifacts. These operations are scoped to the migration and normalization workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 12:24 PM
Security Audit — agent-trust-hub — normalize-skill-for-claude