- Home
- Skills
- Agents & Orchestration
- Agent Session Observatory Lite
Agent Session Observatory Lite
Free lite version, session summary only, no question text, no context percent.
Free
Works with the AI Agents you already use
Agent Session Observatory Lite
Example session with this skill installed
Run a session summary for the last 24 hours to see which Claude Code tasks are waiting for me.
- Read your context and instructions
- Compiled the agent session observatory
I have scanned your local log directories for Claude Code and Codex sessions from the last 24 hours. Below is the summary of active and recently modified sessions.
Agent Session Summary
| Vendor | Project / Title | Status | Model | Last Active |
|---|---|---|---|---|
| Claude | refactor-auth-logic | WAITING | claude-3-7-sonnet | 2 mins ago |
| Claude | unit-test-api-v2 | WORKING | claude-3-7-sonnet | 10 secs ago |
| Codex | schema-migration-sql | IDLE | codex-latest | 5 hours ago |
| Generic | docs-update-workflow | DONE | gpt-4o | 1 hour ago |
| Claude | css-fix-sidebar | WAITING | claude-3-5-sonnet | 15 mins ago |
Next steps
- Resume the
refactor-auth-logicsession to answer the pending architect question. - Check the
css-fix-sidebarsession as it has been waiting for over 15 minutes. - Run
python3 -I session-summary.py --jsonif you need to export this table for your personal dashboard.
Connects securely to your tools. The creator never sees your data.
About this skill
Context switching between multiple local agents is slow and manual. This skill eliminates the friction of hunting through log directories to find which background tasks are finished or stalled. It provides a lightweight, local-only dashboard for developers managing concurrent agent workflows.
What it does
- Status tracking identifies if sessions are currently working, waiting for user input, idle, or completed.
- Multi-agent support natively parses logs from Claude Code and Codex while providing a generic adapter for other .jsonl sources.
- Cross-platform compatibility operates on Mac, Windows, and Linux using only the Python standard library.
- Log aggregation collects session titles, model types, and activity timestamps into a single terminal view.
- Privacy-first reporting truncates sensitive text and shortens home paths without uploading data or making network calls.
How it works
- Run
python3 -I session-summary.pyto get a high-level list of all active and recent agent sessions. - Use the
--hoursflag to filter logs from a specific timeframe, such as the last 72 hours. - Pass the
--jsonflag to output machine-readable data for integration with other local CLI tools. - Point the script at custom log locations using environment variables like
AGENT_OS_LOG_DIRS.
Frameworks & tools
Built with Python 3.11+ using the standard library only. It integrates with Claude Code, Codex, and any agent that exports to .jsonl formats.
Why this beats prompting it yourself
Parsing nested, concurrent .jsonl log files manually is error-prone and tedious. This tool provides a deterministic, zero-dependency script that extracts status logic that a standard LLM prompt cannot access without being fed the entire raw log directory.
Use cases
- Identify which background Claude Code task is waiting for a prompt after you stepped away.
- Audit which models were used across different local projects for cost or performance tracking.
- Monitor long-running Codex sessions to ensure they haven't timed out or gone idle.
Known limitations
Requires Python 3.11 or newer. It cannot display question text or tool call counts as those features are reserved for the full version. Local log access is required.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires Python 3.11+. Does not display session questions, tool call counts, or live status line captures. Local log access is mandatory.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 8 days ago
- Free with an Agensi account
- Software
Python 3.11
Python 3PowerShell