The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted content from legacy repositories. It includes explicit boundary instructions for the agent to treat this content strictly as data and to ignore any directives found within code comments, issue bodies, or documentation, effectively mitigating the risk of indirect prompt injection.
[COMMAND_EXECUTION]: The skill utilizes standard development tools like git, dotnet, and gh to perform repository analysis and environment checks. These operations are restricted to read-only analysis during the research phases and all significant modifications or external operations require explicit user approval via a pt-BR (Portuguese) language gate.
[EXTERNAL_DOWNLOADS]: For repositories provided as URLs, the skill uses git clone to download source code to a temporary directory for analysis. It treats the downloaded content as read-only evidence and provides clear instructions to avoid writing to or executing the cloned codebase, adhering to secure data handling practices.

migration-planner