The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple local shell and Python scripts for project management, including task refreshing (plan-refresh.sh), project health checks (upgrade-check.py), and build orchestration (merge.py).
[INDIRECT_PROMPT_INJECTION]: The skill reads project-local records and documentation to determine piece status and operational modes, which represents a surface for indirect instruction injection.
Ingestion points: Operational parameters and task definitions are read from masterplan.md and records/references/pieces.md (accessed via SKILL.md and scripts/attempt.sh).
Boundary markers: The instructions do not define specific delimiters or warnings for the content read from these markdown records.
Capability inventory: The skill is capable of executing internal scripts, writing build state to the Git directory, and managing GitHub issues through the gh CLI.
Sanitization: The scripts/attempt.sh utility performs numeric validation on task identifiers and utilizes regular expression matching to verify configuration settings in the project plan.

implement