История аудитов

Pure prompt-based skill for Claude Code agent creation. Static findings are false positives: hex color codes misidentified as MD5 hashes, markdown code blocks misidentified as shell backticks, and CLI command examples misidentified as reconnaissance. The skill includes a Python validation script that safely reads user-specified files without network access or credential handling.

Pure prompt-based skill for Claude Code agent creation. Static findings are false positives: hex color codes misidentified as MD5 hashes, markdown code blocks misidentified as shell backticks, and CLI command examples misidentified as reconnaissance. The skill includes a Python validation script that safely reads user-specified files without network access or credential handling.

Pure prompt-based skill with a Python validation script. No network access, no credential handling, no file system overreach. The script only reads user-specified agent files for validation. All capabilities align with the stated purpose of agent creation and validation.

Pure prompt-based skill with a Python validation script. No network access, no credential handling, no file system overreach. The script only reads user-specified agent files for validation. All capabilities align with the stated purpose of agent creation and validation.

Pure prompt-based skill with a Python validation script. No network access, no credential handling, no file system overreach. The script only reads user-specified agent files for validation. All capabilities align with the stated purpose of agent creation and validation.