The Agent Skills Directory

[REMOTE_CODE_EXECUTION] (HIGH): The documentation and checklists in python-overview/README.md, python-basics/README.md, and python-modern-2026/checklist.md recommend the use of curl -LsSf https://astral.sh/uv/install.sh | sh for environment setup. Piped remote execution is a critical risk; although this pertains to a standard industry tool (uv), the execution method is unverifiable. The severity is dropped one level to HIGH as it is associated with the primary purpose of the developer skill.
[EXTERNAL_DOWNLOADS] (HIGH): The skill references and downloads binaries from astral.sh. This domain is not listed as a Trusted External Source, requiring high-severity classification for unverified external references.
[COMMAND_EXECUTION] (LOW): The skill is designed to execute standard development commands using the Bash tool (e.g., python manage.py, pytest, ruff). These operations are consistent with the intended functionality of a developer assistant.
[PROMPT_INJECTION] (LOW): The SKILL.md entry point defines 'Auto-Investigation' rules that automatically ingest untrusted data from local project files. This constitutes an Indirect Prompt Injection surface.
Ingestion points: manage.py, pyproject.toml, requirements.txt, settings/*.py, services/*.py, conftest.py, factories.py, tasks.py, celery.py, and .pre-commit-config.yaml.
Boundary markers: Absent. The skill provides no instructions to wrap these inputs in delimiters or to ignore instructions embedded within the processed data.
Capability inventory: Bash, Write, Edit, Glob, Grep, Task, AskUserQuestion, TodoWrite, and Skill (tools defined in SKILL.md).
Sanitization: Absent. Project files are read directly to understand patterns without content filtering.

faion-python-developer