The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to audit external Python repositories by reading and analyzing source code, configuration files, and CI/CD scripts. This process inherently exposes the agent to untrusted data that could contain malicious instructions meant to override the auditor's persona or influence the final report.
Ingestion points: The agent ingests project contents using Read, Grep, and Glob tools across multiple analysis waves (e.g., repository-inventory, api-design-analysis).
Boundary markers: While the instructions emphasize evidence-based and objective reporting, there is no technical enforcement of delimiters (like XML tags) when project code is interpolated into the agent's context.
Capability inventory: The agent has access to powerful tools including Bash, Write, Edit, and Agent (for spawning sub-agents), which increases the potential impact of a successful injection.
Sanitization: The skill does not implement specific sanitization or filtering for the code it reads before processing it.
[EXTERNAL_DOWNLOADS]: The environment setup phase fetches the uv package manager installation script from an external source.
Evidence: The file references/tool-installer.md contains the command curl -LsSf https://astral.sh/uv/install.sh | sh.
Context: The download originates from the official domain of a well-known Python tooling provider (Astral), which is consistent with the skill's purpose of managing Python environments.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute various commands for project analysis and setup.
Evidence: Use of commands like uv sync, pytest --cov, and directory traversal logic (find ... | wc -l).
Context: These commands are standard for Python development workflows and are used within the scope of generating a technical health report.

python-health-audit