The Agent Skills Directory

[SAFE]: The skill performs read-only analysis of Swift source code to suggest idiomatic improvements. It specifically excludes sensitive directories such as .build, Pods, and agent-specific configuration folders like .claude. No dangerous command execution, data exfiltration, or obfuscation patterns were detected.- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input in the form of local .swift files. While this creates a surface for indirect prompt injection, the agent's capabilities are strictly limited to reading and reporting, with no tool access for writing files or accessing the network, which significantly mitigates the risk. * Ingestion points: Local files identified via Glob: **/*.swift in SKILL.md. * Boundary markers: Not explicitly defined for read operations, though instructions emphasize confirming findings in context. * Capability inventory: Includes Glob, Grep, and Read. No capabilities for file system modification, network communication, or shell execution were identified. * Sanitization: The skill relies on the agent's internal reasoning and explicit safety gates to distinguish between code and instructions.

axiom-swift-simplifier