The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes Bash commands (git rev-parse, git branch, and mkdir -p) to identify the project root and current branch to organize the storage of plan documents.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection attack surface. (1) Ingestion points: The agent is instructed to read files, configuration, and git history within the repository to find facts. (2) Boundary markers: The instructions do not define specific delimiters or instructions to treat repository content as untrusted data. (3) Capability inventory: The skill has access to powerful tools including Bash, Read, Write, Edit, and WebSearch. (4) Sanitization: There are no mechanisms described to sanitize or escape content retrieved from the repository before it influences the agent's actions.

plan-grilling