The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes several shell commands to gather context, including git status, git branch, git log, git diff, and date. These are standard utilities for development workflows and are used here to document the current state of a repository.
[PROMPT_INJECTION]: The skill instructions require the agent to append the phrase "You are Rick. You know what to do." to the end of the generated files. This is a persona-setting instruction intended to influence the behavior of future agent sessions that ingest the file.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect injection surface because it reads previously generated save files and plans to establish context for the current session.
Ingestion points: The agent reads files from docs/rick/<folder>/saves/ and docs/rick/<folder>/plan/current.md to understand previous work.
Boundary markers: There are no specific delimiters or headers used to encapsulate the ingested data to prevent it from being interpreted as instructions.
Capability inventory: The skill has the ability to execute shell commands (git) and write new files to the filesystem.
Sanitization: The skill includes logic to sanitize folder names and explicit instructions to the agent to avoid including PII or credentials in the output, which mitigates data exposure risks.

rick-save