test-driven-development

Pass

Audited by Gen Agent Trust Hub on Aug 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to run terminal commands, specifically npm test, using file paths that may be provided by the environment or user input.
  • [PROMPT_INJECTION]: The skill uses extremely forceful directives to override agent autonomy, such as 'Delete means delete' and 'Stop. That's rationalization,' which instruct the agent to disregard its own reasoning or standard safety behaviors regarding data preservation.
  • [PROMPT_INJECTION]: A surface for indirect prompt injection exists where the agent is expected to interpret and act upon the output of external commands and the contents of various files without explicit sanitization or instruction boundaries.
  • Ingestion points: The agent ingests file paths, source code content, and the output generated by the npm test command.
  • Boundary markers: The skill does not provide delimiters to separate instruction logic from processed data, nor does it warn the agent to ignore instructions embedded in test failures.
  • Capability inventory: The skill assumes the ability to execute shell commands and modify/delete files on the local filesystem.
  • Sanitization: There is no evidence of path validation or output escaping to prevent command injection or instruction hijacking.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 23, 2026, 04:00 AM
Security Audit — agent-trust-hub — test-driven-development