The Agent Skills Directory

[Prompt Injection] (SAFE): The skill contains instructional guidance for an AI assistant to help with TCA development. No attempts to bypass safety filters, override system instructions, or extract system prompts were detected.
[Data Exposure & Exfiltration] (SAFE): No hardcoded credentials (API keys, tokens) or access to sensitive system file paths (e.g., SSH keys, AWS credentials) were found. Networking examples in code snippets use standard Swift patterns (URLSession) for illustrative purposes within an application context.
[Obfuscation] (SAFE): No Base64, zero-width characters, homoglyphs, or other obfuscation techniques were identified in the markdown or code snippets.
[Unverifiable Dependencies & Remote Code Execution] (SAFE): The skill does not include package manager installation commands (npm, pip) or patterns for remote script execution (curl|bash).
[Privilege Escalation] (SAFE): No commands related to privilege escalation (sudo, chmod 777) or system-level configuration changes were detected.
[Persistence Mechanisms] (SAFE): No patterns for establishing persistence (crontab, shell profiles, startup services) were found.
[Metadata Poisoning] (SAFE): Skill metadata (name, description) is accurate, descriptive, and free of hidden instructions.
[Indirect Prompt Injection] (SAFE): The skill is a static reference guide and does not define workflows for ingesting untrusted external data that could lead to indirect injection.
[Time-Delayed / Conditional Attacks] (SAFE): No logic was found that gates behavior based on date, time, or specific environment triggers.
[Dynamic Execution] (SAFE): No use of unsafe dynamic execution (eval, exec) or unsafe deserialization (pickle, marshal) was identified in the provided Swift examples.

composable-architecture