The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local commands to manage Jujutsu repositories and configuration. Key operations include repository initialization with jj git init --colocate, execution of a local Python audit script (scripts/audit_jj_config.py), and configuration updates via jj config set. The skill implements safety measures by displaying commands for user confirmation before execution and handles environment-specific permission requirements for configuration writes by suggesting escalation if standard writes fail.
[PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection by processing local repository configuration data.
Ingestion points: Configuration values are retrieved via jj config list and parsed by scripts/audit_jj_config.py.
Boundary markers: The skill does not employ specific delimiters to isolate external configuration data from the summary instructions.
Capability inventory: The agent has the capability to execute jj commands and modify local configuration files.
Sanitization: No explicit validation or sanitization is performed on the configuration values before they are summarized for the user.
Note: While this ingestion of external data creates a potential injection surface where malicious configuration values could theoretically influence agent behavior, the risk is considered low and associated with the tool's primary purpose of auditing user-controlled settings.

jj-init