The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute a shell command using the kodycodes CLI to automate the filing of friction logs.
Evidence: The skill provides a command block starting with npx @kodycodes/cli execute --local in SKILL.md.
[EXTERNAL_DOWNLOADS]: The execution of the command relies on downloading the @kodycodes/cli and @kentcdodds/friction-log packages from the npm registry.
Evidence: The command includes references to @kodycodes/cli and the package @kentcdodds/friction-log.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process data from external sources, such as repository files and existing issues, which creates a potential surface for indirect prompt injection if those sources contain malicious instructions.
Ingestion points: The skill instructs the agent to search open issues and review 'relevant files' to fill out the friction form.
Boundary markers: No technical delimiters are used for the parameters passed to the command, though instructional warnings to 'Keep secrets and private data out' are present.
Capability inventory: The skill utilizes shell command execution via npx as its primary action.
Sanitization: The skill relies on natural language instructions for the agent to exclude sensitive data rather than programmatic sanitization.

friction-log