The Agent Skills Directory

[SAFE]: The skill defines a structured methodology for translating source code into Allium specifications. It includes clear guidelines for abstraction, specifically directing the agent to strip out implementation-level details like security tokens, secrets, and infrastructure choices (e.g., specific databases or API keys), which reduces the risk of credential exposure in generated documentation.
[SAFE]: The 'orchestration model' for large codebases uses a fan-out pattern with subagents to process data in slices. This approach is designed to manage context window limits and does not involve unauthorized remote code execution or persistence mechanisms.
[SAFE]: No malicious patterns such as prompt injection, obfuscation, or unauthorized data exfiltration were detected. The skill uses standard domain-specific commands for linting and analysis.

distill