The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill utilizes the astro CLI for project initialization, environment management, and DAG parsing. These commands are standard and appropriate for the skill's stated purpose of managing Astro projects.\n- [CREDENTIALS_UNSAFE] (LOW): The airflow_settings.yaml example includes login and password fields with placeholder values. Although the values are generic, providing templates that encourage storing credentials in plaintext project files can lead to accidental exposure if users follow the pattern and commit these files to version control.\n- [EXTERNAL_DOWNLOADS] (LOW): The skill provides a Dockerfile example using an extra-index-url for pip. This mechanism is a known vector for dependency confusion attacks if pointing to untrusted or compromised private registries.\n- [PROMPT_INJECTION] (LOW): (Category 8
Indirect) The skill processes project-level configuration files (requirements.txt, airflow_settings.yaml, Dockerfile, and DAGs). If an agent uses this skill on a repository controlled by an attacker, malicious instructions embedded in these files could influence the agent's behavior during project setup or validation. Evidence: 1. Ingestion points: requirements.txt, packages.txt, airflow_settings.yaml, Dockerfile. 2. Boundary markers: Absent. 3. Capability inventory: astro dev init, astro dev restart, astro dev parse, pip install. 4. Sanitization: Absent.

setting-up-astro-project