The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill includes a reference link to a social media post on X (x.com) belonging to the author (Jamkris) for supplemental architectural guidance. This is a well-known service and the resource is author-owned.
[COMMAND_EXECUTION]: The documentation describes using tools for file retrieval using standard glob patterns (e.g., src/**/*.ts). This is part of the intended pattern for context gathering from a codebase.
[PROMPT_INJECTION]: The skill provides structured instructions for agents to follow when refining queries and evaluating file relevance. These are standard task-oriented guidelines and do not contain safety-filter bypasses or malicious overrides.
[PROMPT_INJECTION]: As the skill involves ingesting content from a codebase, it presents an indirect prompt injection surface.
Ingestion points: Content retrieved from files via the retrieveFiles logic.
Boundary markers: Not explicitly defined in the pattern snippets.
Capability inventory: File reading through retrieval tools.
Sanitization: The pattern focuses on relevance scoring rather than input sanitization.

iterative-retrieval