by dreadnode · Agent Tool · ★ 75
AgentLens Developed at MATS Exploration Phase under Neel Nanda, for a research project with Greg Kocher. A harness for running multi-session agent trajectories using the Claude Agent SDK, capturing them in ATIF (Agent Trajectory Interchange Format), and tracking file state changes across sessions. Built for AI alignment and interpretability research — studying how LLM agents behave across multi-turn, multi-session, multi-agent interactions. Note: AgentLens currently supports Claude Code via the Claude Agent SDK. Support for additional agents and frameworks is planned — see Roadmap.
| Stars | 75 |
| Forks | 6 |
| Language | Python |
| Category | Agent Tool |
| License | MIT |
| Quality Score | 65.096/100 |
| Open Issues | 1 |
| Last Updated | 2026-03-19 |
| Created | 2026-03-18 |
| Platforms | python |
| Est. Tokens | ~138k |
These tools work well together with agent-lens for enhanced workflows:
Explore other popular agent tool tools:
agent-lens is Agent observability and replay tooling for AI safety & interpretability research.. It is categorized as a Agent Tool with 75 GitHub stars.
agent-lens is primarily written in Python. It covers topics such as ai, alignment, interpretability.
You can find installation instructions and usage details in the agent-lens GitHub repository at github.com/dreadnode/agent-lens. The project has 75 stars and 6 forks, indicating an active community.
agent-lens is released under the MIT license, making it free to use and modify according to the license terms.