The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill involves ingesting and transforming existing scripts and documentation which may contain untrusted data or malicious instructions.
Ingestion points: The skill processes shell scripts, Python glue code, Makefiles, CI job configurations, and prompt chains found in documentation (SKILL.md).
Boundary markers: There are no explicit instructions or delimiters defined to isolate the source code being analyzed from the agent's internal instructions.
Capability inventory: The migration targets a framework capable of network operations (nika:fetch), file system access (nika:read, nika:write, nika:edit), and shell command execution (exec:, shell:).
Sanitization: The instructions do not specify a sanitization or validation step for the content of the legacy scripts before they are converted into executable workflows.
[COMMAND_EXECUTION]: The skill provides documentation on how to translate existing shell commands and complex pipelines into a new workflow format.
Evidence: The mapping table in SKILL.md explicitly shows how to handle exec: and shell: commands for tasks that cannot be handled by native tools.

nika-migration