More screenshots
Works with the AI Agents you already use
Session Finder
Search, read, and resume local Claude Code and Codex CLI conversation logs across all projects.
$5
Session Finder
Example session with this skill installed
Find the session from the last 2 days where I worked on the auth system in the backend project, then show me the transcript.
- Read your context and instructions
- Compiled the session finder
I have searched your local logs for sessions related to the auth system in the backend project from the last two days. I found one matching session and retrieved the transcript.
Session Search Results
| Session ID | Agent | Project | Title | Tokens | Resume Command |
|---|---|---|---|---|---|
a1b2c3d4 | Claude | backend | Refactor JWT auth flow | 4,250 | claude --session a1b2c3d4 |
Transcript: a1b2c3d4 (Refactor JWT auth flow)
User: We need to update the token expiration logic in src/auth.ts.
Agent: I will check the current implementation.
- Tool Call: read_file("src/auth.ts")
Agent: I see the expiration is hardcoded to 1h. I'll move this to an environment variable. - Tool Call: write_to_file("src/auth.ts", ...)
User: Also update the test suite to match. - Tool Call: ls("tests/")
- Tool Call: run_command("npm test")
Agent: The tests passed. The auth logic now usesprocess.env.JWT_EXPIRY.
Next steps
- Run
claude --session a1b2c3d4in your terminal to continue this conversation. - Use
python3 scripts/sessions.py show a1b2c3d4 --last 200if you need the full source code diffs from that session.
Connects securely to your tools. The creator never sees your data.
About this skill
Stop losing context between agent sessions. This skill enables deep searching and retrieval across local Claude Code and Codex CLI logs, allowing you to recover specific conversations, commands, and logic without manual file crawling.
What it does
- Multi-agent log retrieval reads local session histories from both Claude Code and Codex CLI.
- Deep search queries session titles, folder paths, git branches, and user prompts.
- Transcript recovery extracts conversation history and tool execution sequences from specific session IDs.
- Usage tracking calculates input and output token counts for individual sessions.
- Workflow resumption provides the exact CLI commands needed to resume or fork any historical session.
How it works
- List sessions using the
listcommand to see recent activity filtered by date, project, or agent type. - Search history via the
searchcommand to locate specific work, such as a bug fix or a specific prompt. - Inspect transcripts using the
showcommand with a session ID to review what was discussed and which tools were called. - Resume work by copying the provided resume or fork command into a new terminal session.
Frameworks & tools
Requires Python 3.10+ and zstd for older Codex logs. It interacts directly with ~/.claude/projects/ and ~/.codex/sessions/ directories.
Why this beats prompting it yourself
Standard agents cannot see their own past session logs unless they are currently active in them. This skill provides the specific filesystem logic and decompression handlers required to index, search, and parse the proprietary log formats of both major CLI agents.
Use cases
- Context recovery to find how a specific login bug was fixed three days ago.
- Token auditing to identify which projects or sessions are consuming the most resources.
- Session migration to generate the fork command for moving an old conversation into a new branch.
- Knowledge retrieval to summarize technical decisions made in a previous project session.
Known limitations
Search only covers user prompts and session metadata, not agent responses. Codex logs older than one week require zstd to be installed on the system path.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for LLM Ops.
Requires Python 3.10+. Requires zstd for Codex logs older than 7 days. Does not search agent replies, only user prompts and metadata. Output tokens exclude cache hits.
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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 10 days ago