The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a PowerShell script (scripts/run_forge_audit.ps1) designed to locate and execute the skyrim_forge Python module to perform plugin diagnostics.
[COMMAND_EXECUTION]: The test suite (scripts/test_plugin_header_fixtures.py) uses the subprocess module to execute the audit_plugin_headers.py script against dynamically generated test fixtures.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted binary data from Skyrim plugin files and archives, which could potentially contain malicious instructions embedded in record names or metadata that are then displayed to the agent.
Ingestion points: The scripts/audit_plugin_headers.py script reads bytes from .esp, .esm, and .esl files, including those contained within ZIP archives.
Boundary markers: The output generated by the audit scripts is not wrapped in specific delimiters to prevent the agent from interpreting extracted strings as instructions.
Capability inventory: The skill has the capability to execute shell commands via the skyrim_forge wrapper script (scripts/run_forge_audit.ps1) and the test script (scripts/test_plugin_header_fixtures.py).
Sanitization: The parser extracts record signatures and master names using latin1 or cp1252 decoding without sanitizing the resulting strings before printing them to the console.

skyrim-plugin-authoring