android-expert

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its core function of reviewing untrusted code.
  • Ingestion points: Processes user-provided Android code, project configurations, and documentation (SKILL.md instructions).
  • Boundary markers: No explicit delimiters or instructions are provided to ignore embedded commands in the data being reviewed.
  • Capability inventory: Access to Bash, Write, Read, Edit, Grep, and Glob tools (SKILL.md tools field).
  • Sanitization: No sanitization or validation of the input code is defined.
  • [EXTERNAL_DOWNLOADS]: The metadata includes an installation command that fetches content from an external repository.
  • Evidence: curl -sL "https://raw.githubusercontent.com/majiayu000/claude-skill-registry-data/main/data/android-expert/SKILL.md" in metadata.json.
  • [COMMAND_EXECUTION]: The skill instructions define a 'Memory Protocol' that executes shell commands to maintain context across sessions.
  • Evidence: cat .claude/context/memory/learnings.md in SKILL.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 04:37 PM
Security Audit — agent-trust-hub — android-expert