The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided Swift source code and compiler diagnostic messages to provide remediation advice, creating a potential surface for indirect prompt injection via embedded instructions in comments or string literals.\n
Ingestion points: The 'Diagnostic Workflow' in SKILL.md and the fix guides in references/diagnostics.md ingest external diagnostic and code text.\n
Boundary markers: The skill does not provide instructions to use delimiters or specific safety wrappers when processing these external inputs.\n
Capability inventory: The skill has no functional scripts, tool definitions, or permissions to perform network or file system operations; it is entirely documentation-based.\n
Sanitization: No procedures for sanitizing or escaping user-provided text are specified.\n- [EXTERNAL_DOWNLOADS]: The skill references the apple/swift-async-algorithms library as a recommended dependency for project implementation.\n
Source: The reference points to a repository under the trusted apple organization on GitHub.\n
Context: It is presented as a standard and safe resource for implementing AsyncSequence patterns.

swift-concurrency