The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill contains numerous references to official documentation and package registries.
Links to official documentation for libraries including httpx, GitPython, Arrow, Fabric, and Prefect.
References to authoritative Python documentation sites such as peps.python.org and packaging.python.org for specification research.
Redirections to the official Astral domain (astral.sh) for tool installation.
[REMOTE_CODE_EXECUTION]: Documentation within the tool registry describes the standard installation method for the uv package manager using a piped shell script from its official site.
Evidence: curl -LsSf https://astral.sh/uv/install.sh | sh in references/tool-library-registry.md.
[COMMAND_EXECUTION]: The skill provides build hook templates that execute local scripts during the package build process.
Evidence: assets/hatch_build.py includes logic to run scripts/build-binaries.sh or scripts/build-binaries.py using subprocess.run with list-based arguments. This is a legitimate pattern for the Hatch build backend to handle compilation steps.
[INDIRECT_PROMPT_INJECTION]: The skill's architectural design involves automated analysis of user-provided codebases, which creates an ingestion surface for untrusted data.
Ingestion points: The skill reads project source files, pyproject.toml, and git logs (git log --oneline) to gather context for planning and review (found in references/python-development-orchestration.md).
Boundary markers: Missing explicit delimiters or warnings to ignore instructions within ingested codebase data.
Capability inventory: The skill uses tools capable of file system writes and package installations (uv, hatchling).
Sanitization: The skill relies on standard static analysis tools (ruff, mypy) to validate the resulting output.

python3-development