The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/archive_todo.py executes a fixed shell command git rev-parse --show-toplevel to identify the repository root. This is a standard development operation and does not involve user-controlled input in the command execution path.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes the contents of TODO.md to archive tasks. This presents a surface for indirect prompt injection if the task descriptions contain malicious instructions intended to influence the agent's behavior during the archiving process or in subsequent steps where the agent reads the archive.\n
Ingestion points: scripts/archive_todo.py reads the entire content of <root>/TODO.md.\n
Boundary markers: None. The skill does not implement delimiters or safety instructions when processing the content of tasks.\n
Capability inventory: The skill has the capability to write to the local filesystem (TODO.md and .ai/todos/ directories) and execute the git binary via subprocess.\n
Sanitization: The script performs Markdown parsing to identify tasks and headings but does not sanitize or filter the actual text content of the tasks being archived.

todo-archive