Version updated for https://github.com/Siddhant-K-code/agent-trace to version v0.94.1. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary agent-trace is a tool that captures and inspects the actions of AI agents, providing insights into file operations, tool calls, decision points, error recovery, and commands executed. It helps automate debugging by replaying sessions and exporting data to various observability platforms like Datadog, Honeycomb, New Relic, or Splunk. The action is experimental and requires careful consideration before use in production environments.

Version updated for https://github.com/Siddhant-K-code/agent-trace to version v0.93.1. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary agent-trace is an experimental tool designed to capture and replay AI agent sessions, providing insights into file operations, tool calls, decision points, error recovery, and commands executed by the agent. It helps developers understand and audit AI-assisted work after the fact, ensuring transparency and accountability. The tool can be installed via pip and offers CLI hooks, an MCP proxy, and a Python decorator for integration with various coding tools.

Version updated for https://github.com/Siddhant-K-code/agent-trace to version v0.93.0. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary The agent-trace action captures and records AI agent sessions, providing detailed information about what files were read, tool calls made, changes and failures during execution. It enables replaying these sessions later, facilitating inspection of the agent’s behavior and debugging issues after a pull request is merged. The action supports exporting data to various observability platforms like Datadog, Honeycomb, New Relic, or Splunk for monitoring purposes and includes features to set rules to control the agent’s execution, such as cost limits and file access restrictions.

Version updated for https://github.com/Siddhant-K-code/agent-trace to version v0.87.0. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary agent-strace is a tool designed to capture and replay the full session of AI agents, providing insights into tool calls, file operations, decision points, and error recovery. It helps in inspecting what ran, what changed, what failed, and where human review should occur, making it easier to debug and audit AI-assisted engineering processes.

Version updated for https://github.com/Siddhant-K-code/agent-trace to version v0.83.0. This action is used across all versions by ? repositories. Action Type This is a Composite action. Go to the GitHub Marketplace to find the latest changes. Action Summary The agent-trace GitHub Action is designed to trace and replay the actions of AI agents, providing insights into tool calls, file operations, decision points, error recovery, and actual commands. It captures the full session and allows for exporting to observability tools like Datadog, Honeycomb, New Relic, or Splunk, setting rules to control the agent’s behavior. The action supports CLI hooks, MCP proxy integration, and Python decorator options for capturing tool calls without needing an MCP server.