The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze execution journals (traces) from the .nika/traces/ directory to diagnose failures.
Ingestion points: Data from external trace files is loaded into the agent's context using commands such as nika trace show, nika trace replay, nika trace outputs, and nika trace peek (SKILL.md).
Boundary markers: There are no explicit instructions to treat trace content as untrusted or to use specific delimiters when processing the findings.
Capability inventory: The skill utilizes the nika CLI for filesystem access, verification of hashes, and workflow execution/testing.
Sanitization: The instructions do not include specific sanitization or validation steps for the content found within the trace journals.
[COMMAND_EXECUTION]: The skill's primary functionality is delivered through the execution of various nika CLI subcommands for forensics, verification, and run management.
[EXTERNAL_DOWNLOADS]: The documentation includes instructions to install the nika tool suite using Homebrew from the author's own repository (supernovae-st/tap/nika).

nika-debugging