История аудитов

Static command, PowerShell, sensitive-data, weak-crypto, and reconnaissance detections were mostly Markdown or keyword false positives in SKILL.md, not executable source code. However, the skill contains a high-risk instruction to treat a system_instructions block from task JSON as hard constraints, which can let untrusted task data override the agent. The skill also instructs agents to run local workflow scripts that can write, append, flag, or move evidence files, so publication should require remediation and sandbox guidance.

Static scanner flagged patterns related to external commands and file operations. Evaluation confirms these are legitimate document processing workflows. No malicious intent found. The skill runs hardcoded Python scripts for fetching, analyzing, and submitting legal document analysis tasks. All commands are predefined workflow operations with no user input injection risk.

AI analysis failed after multiple attempts - MANUAL REVIEW REQUIRED before publishing. This skill cannot be auto-published until reviewed by a human.

AI analysis failed after multiple attempts - MANUAL REVIEW REQUIRED before publishing. This skill cannot be auto-published until reviewed by a human.

Prompt-only skill containing AI instructions for document analysis. No executable code, no direct filesystem access, no network calls. External scripts referenced but not included. Pure instruction-based workflow for legal document processing.

Prompt-only skill containing AI instructions for document analysis. No executable code, no direct filesystem access, no network calls. External scripts referenced but not included. Pure instruction-based workflow for legal document processing.

Prompt-only skill containing AI instructions for document analysis. No executable code, no direct filesystem access, no network calls. External scripts referenced but not included. Pure instruction-based workflow for legal document processing.