Session Lens — read your AI agent's session file, then review the run

Drop the .jsonl session file your coding agent just wrote and finally read it: a turn-grouped chat view with thinking blocks, tool calls and inline edit diffs, a Gantt timeline of tool-call durations, an exact token and cost breakdown at current list price, and a census of every failure, repeated identical call and file touched — all computed in your browser, with nothing uploaded. Then ask a model about it in one of three lanes: review the run (where it wasted effort), write the handoff (what it actually changed, as a PR note), or find the root cause (why it failed). Every count the review states is checked back against the arithmetic of your own file, and the disagreements are printed above the findings — where the review and the file disagree, the file wins. Reads Claude Code, OpenClaw and Gemini CLI session formats. Session parsing and the viewer are derived from the open-source agent-chatlens project by kangjinghang (MIT licence); this app is an independent derived work and is not affiliated with that project, Anthropic, OpenClaw or Google.

Every public app is built from a security-scanned skill and must pass a clean scan — skill and frontend — before it can be listed. Have a skill of your own? Turn it into an app — or read the step-by-step walkthrough.