The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill contains no executable code or scripts. The compatibility metadata explicitly states there are no runtime dependencies, and no remote fetching patterns (like curl|bash) were detected.
[DATA_EXFILTRATION]: No network operations or commands to read sensitive files (such as credentials, SSH keys, or environment variables) were found. The skill correctly identifies secret management as a specialized concern for security engineering.
[PROMPT_INJECTION]: The instructions are purely technical and architectural in nature. There are no attempts to override agent behavior, bypass safety guardrails, or use role-play/DAN injection techniques.
[OBFUSCATION]: All content is provided in clear text. There is no evidence of Base64 encoding, zero-width characters, homoglyphs, or hidden strings designed to evade detection.
[COMMAND_EXECUTION]: The skill does not invoke shell commands or perform subprocess spawning. It focuses on design and implementation workflow guidance for mobile development.
[INDIRECT_PROMPT_INJECTION]: The skill does not provide mechanisms for ingesting untrusted external data into the agent's context. While it discusses mobile bridge input, it specifically advises validating messages and versioning contracts to prevent unrestricted capability exposure.
[DYNAMIC_CONTEXT_INJECTION]: No use of the dynamic command execution syntax (!command) was found in the SKILL.md file or any referenced documents.

mobile-engineering