The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes git log to provide commit volume statistics. This is performed via a Bash environment restricted to git-prefixed commands.
[COMMAND_EXECUTION]: The skill references a beans CLI tool to fetch incident data from a task backend. This tool is used for local data aggregation related to the retrospective summary.
[PROMPT_INJECTION]: The skill contains an indirect prompt injection surface (Category 8) as it reads and summarizes content from local markdown files (learned/*.md) and task backend outputs.
Ingestion points: Markdown files in the .groove/memory/learned/ directory and output from the beans CLI tool.
Boundary markers: The skill does not employ specific delimiters or instructions to ignore instructions that might be embedded in the data it summarizes.
Capability inventory: The skill is restricted to Read, Glob, and Bash(git:*) tools, with no permissions for file writing or network access, which limits the potential impact of an injection.
Sanitization: The skill performs structural parsing of data tables but does not sanitize the descriptive text from learnings or incidents before outputting them to the conversation.

groove-utilities-memory-retrospective