The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection due to the ingestion of untrusted external data.
Ingestion points: Processes local dependency manifests (requirements.txt, pyproject.toml, Pipfile) and retrieves content from external sources via WebSearch and documentation MCP tools.
Boundary markers: No explicit delimiters or instructions are provided to isolate untrusted data from the agent's instruction context.
Capability inventory: The agent executes shell commands for package auditing, upgrading, and verification (pip, poetry, pipenv, pytest) as defined in SKILL.md and references/python_venv_handling.md.
Sanitization: The skill lacks specified validation or sanitization routines for the data ingested from dependency manifests or web results before it is used in logic or commands.
[COMMAND_EXECUTION]: The skill performs shell command execution to fulfill its intended purpose of managing Python environments and dependencies.
Evidence: Phases 2 through 5 in SKILL.md utilize commands like pip install, poetry update, pipenv check, and pytest for installation and verification.
[EXTERNAL_DOWNLOADS]: The skill identifies a remote source for fetching shared configuration files if they are missing locally.
Evidence: SKILL.md provides a fallback mechanism using WebFetch to retrieve files from the author's repository at https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/skills/.

ln-823-pip-upgrader