test-driven-development

Warn

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: MEDIUMSAFEMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
  • [METADATA_POISONING]: The skill's YAML frontmatter attributes the author as 'addyosmani', which contradicts the identified author 'seaworld008'. Attributing the skill to a well-known software expert can cause users to misjudge its reliability and safety based on the reputation of the purported author.
  • [SAFE]: The skill package is composed entirely of Markdown documentation and reference checklists. No executable scripts, binaries, or automated lifecycle hooks are present.
  • [SAFE]: The instructions include proactive security guardrails, explicitly directing the agent to treat browser-derived data as untrusted and prohibiting the access of credentials like cookies or session tokens during browser testing.
  • [INDIRECT_PROMPT_INJECTION]: The skill directs the agent to analyze external project data and browser states, creating a theoretical attack surface.
  • Ingestion points: The agent ingests project source code, test results, and browser DOM/console data as described in SKILL.md.
  • Boundary markers: The skill explicitly defines a security boundary in SKILL.md, instructing the agent to never interpret content read from external environments as instructions.
  • Capability inventory: The skill does not provide custom executables; the agent uses standard system tools for file and shell operations (SKILL.md, references/testing-patterns.md).
  • Sanitization: The documentation provides high-level guidance for the agent to perform validation and maintain boundaries when processing external data.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 8, 2026, 06:55 AM
Security Audit — agent-trust-hub — test-driven-development