The Agent Skills Directory

[SAFE]: Comprehensive analysis of the provided Python scripts and documentation confirms that the tool operates locally using standard libraries. No malicious patterns such as credential harvesting, data exfiltration, or unauthorized persistence mechanisms were identified.
[EXTERNAL_DOWNLOADS]: The documentation references external skills for PDF and Word document generation from a trusted organization. These are suggested integrations for report formatting and do not involve the automatic execution of untrusted remote code.
[INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection because it ingests and processes untrusted C-AL source code files. If these files contain malicious instructions, an agent reading the generated report could potentially be influenced.
Ingestion points: scripts/parser.py reads .cs and .c-al source files from the workspace.
Boundary markers: Absent; analysis findings and code snippets are presented to the agent without explicit delimiters.
Capability inventory: The agent executes local scripts and possesses the ability to read and write report files.
Sanitization: None; the tool uses regex-based patterns to extract data from raw source code.

codeunit-analyzer