The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted content from external Git repositories, which serves as a vector for indirect prompt injection attacks. \n
Ingestion points: The agent retrieves data from external repositories via get_file_contents, search_code, list_commits, and get_commit as described in SKILL.md and references/history-investigation.md. \n
Boundary markers: The instructions do not require the agent to use delimiters or specific boundary markers when processing these external inputs to distinguish between instructions and data. \n
Capability inventory: The skill utilizes a suite of GitHub exploration tools including get_repository_tree, get_file_contents, search_code, list_branches, list_tags, list_commits, get_commit, and search_commits. \n
Sanitization: There is no evidence of sanitization or validation of the data retrieved from the repositories before it is processed by the agent.

codebase-exploration