The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Jujutsu (jj) version control system to gather context. It executes commands such as jj diff -s and jj diff -- path to analyze code changes.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via external data sources.
Ingestion points: The skill ingests untrusted data from the web using WebFetch and processes local code via jj diff output.
Boundary markers: No specific delimiters or instructions are provided to the agent to distinguish between its own instructions and the content being processed.
Capability inventory: The skill can execute shell commands (jj) and spawn additional sub-agents via the Task tool.
Sanitization: There is no evidence of sanitization or validation of the ingested content before it is processed or passed to other agents.

code-refactor