dejadoc

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute 'cargo dejadoc' and its variations (e.g., --json, -v) to identify duplicate doctests.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of Rust source code and tool output to make decisions about code modifications.
  • Ingestion points: Rust source files containing doctests and the JSON output of the 'cargo dejadoc' command.
  • Boundary markers: None identified; instructions do not specify delimiters for external content.
  • Capability inventory: The skill has the capability to modify, delete, and rewrite blocks of code within the file system.
  • Sanitization: No explicit sanitization or validation of the ingested code blocks or tool output is described.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 06:57 PM
Security Audit — agent-trust-hub — dejadoc