The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a wrapper script (scripts/run.py) to execute the maintenance engine via uv. This script uses subprocess.run and os.execvpe to launch the engine as a separate process.
[EXTERNAL_DOWNLOADS]: The skill installs the exmergo-dex-core package from PyPI. The version is strictly pinned to 1.17.3 in the wrapper script, ensuring consistent and predictable deployment.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data that could contain malicious instructions.
Ingestion points: Data enters the context via database warehouse comments, SQL files, YAML configurations, and ticket/issue text.
Boundary markers: The skill contains explicit instructions in SKILL.md (Spending authority section) telling the agent to treat these sources strictly as data.
Capability inventory: The primary capability is the execution of the maintenance engine via scripts/run.py which can read from and propose changes to the project.
Sanitization: The skill relies on the agent's adherence to the "Spending authority" rules, which forbid data from authorizing sensitive actions like bypassing budgets or confirming destructive operations.

maintain