The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes git status and git diff --stat to programmatically identify modified files in the repository. These commands are restricted to local repository state and are standard for development workflows.
[DATA_EXPOSURE]: The skill processes content from plan.md and todo.md to summarize project progress. This data exposure is localized to the project environment and serves the intended purpose of the skill.
[INDIRECT_PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it interpolates data from potentially untrusted files into its output.
Ingestion points: Files plan.md and todo.md are read during Step 2.
Boundary markers: Absent. The gathered text is placed directly into markdown templates without delimiters.
Capability inventory: File system write access (creating markdown files) and subprocess execution (git).
Sanitization: None. The content from the project files is used directly as narrative text in the handover document.

generate-handover