The Agent Skills Directory

[COMMAND_EXECUTION]: The system assessment template in references/system-assessment.md includes instructions to execute shell commands via Python's subprocess module.
Evidence: subprocess.run(['git', 'log', '--format=format:', '--name-only'], ...) used to identify development hotspots in the legacy codebase.
Evidence: Mention of using dot (GraphViz) via command line to render dependency graphs.
[EXTERNAL_DOWNLOADS]: The skill references and recommends the installation of various third-party development, testing, and security libraries.
Evidence: Recommends pip install mutmut for mutation testing in references/legacy-testing.md.
Evidence: Mentions and utilizes libraries such as pytest, pytest-cov, approvaltests, syrupy, hypothesis, six, httpx, and sqlalchemy across multiple reference files.
Evidence: Suggests using security and analysis tools including pylint, radon, sonarqube, bandit, and safety in references/system-assessment.md.
[INDIRECT_PROMPT_INJECTION]: The skill provides templates for an automated codebase analyzer (LegacyCodeAnalyzer), which creates an attack surface for indirect prompt injection from analyzed source code.
Ingestion points: The LegacyCodeAnalyzer class in references/system-assessment.md reads and parses all Python files (*.py) within a user-provided directory path.
Boundary markers: The template does not include specific boundary markers or sanitization logic to prevent malicious comments or code patterns in the analyzed files from influencing the agent's summary.
Capability inventory: The analyzer templates use ast.parse and subprocess.run (for Git logs), providing a pathway for code to influence system analysis.
Sanitization: No explicit sanitization or escaping of the content read from files is implemented in the provided Python templates.

legacy-modernizer