v0.1.0 — Initial Release

Changelog

All notable changes to Agent Leak will be documented in this file.

The format is based on Keep a Changelog,
and this project adheres to Semantic Versioning.

[0.1.0] - 2026-10-26

Added

• Core Secret Engine: Multi-signal credential detection (OpenAI, Anthropic, AWS, GitHub PATs, GitLab, Google, Stripe, Slack, Discord, JWT, Bearer tokens, database connection strings, SSH private keys).
• AI Agent Specific Audit (agent-leak agent): Detection of leaked session states (.claude/, .cursor/, .codex/), debug logs, copied terminal outputs, prompt dumps, and unignored sensitive files.
• Git Diff Scanner (agent-leak diff): Staged and unstaged working-tree change inspection with exact line reporting.
• Git History Audit (agent-leak history): Commit log audit distinguishing current vs historical exposures with immediate rotation guidance.
• Pre-Commit Hook (agent-leak install-hook): Native Git hook preventing commit of secrets with documented AGENT_LEAK_ALLOW=1 emergency bypass.
• Privacy Non-Leak Guarantee: Strict masking and redaction across Console, JSON, and SARIF output formats.
• Doctor Command (agent-leak doctor): Runtime environment, Git configuration, and rule engine health check.
• Agent Skill: Operational skill definition for AI coding agents (skill/agent-leak/SKILL.md).
• Interactive Demo: Built-in fake vulnerable demo project under demo/demo-project.