The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses find, ls, and grep commands to discover existing plan files on disk based on a ticket ID or recent activity. These are standard search operations used for auto-discovery.
[REMOTE_CODE_EXECUTION]: The skill defines sub-agents (codebase-locator, codebase-analyzer, codebase-pattern-finder) that use standard tools like Grep, Glob, and Bash to understand the local codebase. These sub-agents are strictly instructed to be 'documentarians' and not to modify code or suggest changes unless requested.
[DATA_EXPOSURE]: The skill accesses implementation plans stored in thoughts/shared/plans/. It includes logic to read these files fully to understand requirements and update them. This is the primary function of the skill and does not involve sensitive system files or credentials.
[DYNAMIC_CONTEXT_INJECTION]: The skill uses $ARGUMENTS to extract a TICKET_ID from the user's request via shell commands. This is used for file path construction and discovery, which is a common pattern in development-oriented AI agent skills.

iterate-plan