Skip to content

Weekly agent setup review from session logs

Every Monday a local job condenses the past week's coding agent transcripts into a digest, then a read only session suggests setup changes such as hooks, skills, permission entries and instruction rules. It writes a report and changes nothing.

EngineeringWorks with Claude Code · CursorCommunity example

What it does

  • Condenses the last 7 days of session transcripts into a small digest of prompts, tool counts, commands, errors and corrections
  • Runs a headless review with read only tools and a spending cap
  • Looks for stale hook paths, repeated manual flows, recurring shell errors and prompts you keep correcting
  • Writes a dated report with evidence for each suggestion
  • Shows a desktop notification when the report is ready

Your approval boundaries

Creates drafts

These actions always wait for your yes:

  • Applying any suggested hook, skill, permission or rule change

You decide what happens next.

More about this workflow

This setup is for developers and small engineering teams who want fewer chores. It takes a repeat chore off your plate and hands back something you can check quickly.

First, it condenses the last 7 days of session transcripts into a small digest of prompts, tool counts, commands, errors and corrections. Next, it runs a headless review with read only tools and a spending cap. Then it looks for stale hook paths, repeated manual flows, recurring shell errors and prompts you keep correcting. After that, it writes a dated report with evidence for each suggestion. Finally, it shows a desktop notification when the report is ready. It works with your docs, each on its own login with limited access.

You get the prep work done and waiting for you. The agent stops and asks for your yes before applying any suggested hook, skill, permission or rule change, and says plainly if a step cannot be undone.

“It only proposes changes.”
Jaelyn chialin6 on GitHub · Oct 14, 2026
Before you startDocs accessScoped permissionsClaude Code or Cursor
Explore required tools and model recommendations

Tools and access

  • Google Docs

    Handles the docs part of this job. Grant one folder, create and edit only.

    API required: Yes. Needs a Google Cloud OAuth app with the Docs API turned on.

Official links only. Not sponsored, no affiliate links.

Model recommendations

Best pick

Claude Opus 5.5

Its maker points to it for long running coding work.

$4 per million input tokens, $20 per million output tokens.

Anthropic docs as of October 11, 2026

Cheaper backup

Claude Sonnet 5.5

Half the price per token and still strong for reviews and docs.

$2 per million input tokens, $10 per million output tokens.

Anthropic docs as of October 11, 2026

Our own picks. Not sponsored, no affiliate links.

Build your prompt

Add your details, then copy the prompt into your tool.

1Your details

Fill these in and the prompt updates as you type.

where your agent sessions are stored

e.g. 7

2Your prompt

4 of 4 still empty
Fill these in before you run it: [TRANSCRIPT FOLDER: where your agent sessions are stored] [DAYS TO REVIEW: e.g. 7] [SPENDING CAP PER RUN] [RUN DAY AND TIME] You are the Weekly agent setup review from session logs agent. Every Monday a local job condenses the past week's coding agent transcripts into a digest, then a read only session suggests setup changes such as hooks, skills, permission entries and instruction rules. It writes a report and changes nothing. Use only the details filled in above. If any [BRACKETED] item is still blank or unclear, stop and ask for it first. Do not guess. What you do on your own: - Condenses the last 7 days of session transcripts into a small digest of prompts, tool counts, commands, errors and corrections - Runs a headless review with read only tools and a spending cap - Looks for stale hook paths, repeated manual flows, recurring shell errors and prompts you keep correcting - Writes a dated report with evidence for each suggestion - Shows a desktop notification when the report is ready What you must never do without a yes from a person: - Applying any suggested hook, skill, permission or rule change Rules: - Before any step on the list above, stop and ask. Say in plain words what you want to do, who it affects, and whether it can be undone. - If something cannot be undone, say so first, in one short line. - Write drafts, not final actions. A person sends, pays, posts, or deletes. - Keep a short log of every action you took and every source you read. - If you are unsure, ask instead of guessing.
View run instructionsPaste into: Claude Code, Cursor

Your answers stay in this browser only.

Run your workflow

Connect your tools with only the access this setup needs, then follow these steps.

Steps

Do these in order. Test on old data before it touches live work.

  1. 1Connect Google Docs with its own login, not yours. Grant one folder, create and edit only.
  2. 2Paste the prompt where your harness takes instructions. See the note under the Copy button.
  3. 3Do a dry run on old or test data first. Check every draft and every stop point by hand before you let it near live work.
  4. 4Read its log after the first real run. Only widen access once a week of runs looks clean.

Safeguards and trust score

The score describes this example setup’s access, approval boundaries, ability to undo changes, and scope.

Access24 / 30

Writes drafts and notes inside your tools, nothing leaves on its own.

Human approval27 / 30

Waits for a yes on: applying any suggested hook, skill, permission or rule change.

Undo and logs14 / 20

Proposes changes to live records, so the setup needs a change log and a way back.

Scope20 / 20

Runs on its own login across 1 tool. Fewer connections means a smaller blast radius.

How to raise it

  • Human approval: Add a second named stop next to "applying any suggested hook, skill, permission or rule change", so every step that leaves your hands waits for a yes, and have the confirm say plainly what cannot be undone +3
  • Undo and logs: Stop it from editing live records. Have it write the change as a proposal you apply, and keep a log of every change with a way to roll it back +3

Do all of these to reach 91, up from 85.

Want help getting there? Get an AI Trust Audit from Pixeldesigns

Reviewed by Trustcase (example setup). How scoring works

Embed badge

Show this trust score on your site or README. It links back here.

Trustcase trust score badge
Markdown
[![Trustcase trust score for Weekly agent setup review from session logs](https://trustcase.dev/badge/weekly-agent-setup-review-from-session-logs.svg)](https://trustcase.dev/use-cases/weekly-agent-setup-review-from-session-logs)
HTML
<a href="https://trustcase.dev/use-cases/weekly-agent-setup-review-from-session-logs"><img src="https://trustcase.dev/badge/weekly-agent-setup-review-from-session-logs.svg" alt="Trustcase trust score for Weekly agent setup review from session logs" height="20"></a>
  • Weekly open source contribution scout

    Once a week it checks the status of your open issues and pull requests in other people's repos, finds unclaimed help wanted issues in the projects you care about, and writes one issue in your own repo with what needs you. It posts nothing anywhere else.

  • Pull request reviewer

    Gives every PR a first pass before a person looks.

  • Flaky test hunter

    Finds tests that fail at random and why.

  • Incident timeline

    Builds the timeline during an outage so people can focus on the fix.

Other workflows that use the same tools