The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill utilizes a shell command (git log) within the context metadata field to retrieve recent commit history at load time. This implementation is safe as it does not incorporate user-supplied arguments, access sensitive credentials, or perform network requests.
[COMMAND_EXECUTION]: The skill instructions specify the use of git log to analyze development history. This is a legitimate and expected operation for the skill's primary purpose of generating a retrospective.
[DATA_INGESTION_SURFACE]: The skill reads from sprint plans, milestone definitions, and codebase comments (TODO/FIXME). While these represent ingestion points for potentially untrusted data that could host indirect prompt injections, the skill's logic is restricted to data extraction and metrics calculation, which mitigates the risk of the agent following embedded instructions.

retrospective