CtrlK
BlogDocsLog inGet started
Tessl Logo

markusdowne/memory-roundtrip-guard

Tests memory writes, confirms read-back accuracy, and validates retrieval success to ensure saved information can actually be recovered. Use when you need to verify memory was saved correctly, check if stored data can be retrieved, confirm a memory entry is discoverable, or escalate when saved information appears lost or corrupted. Covers write confirmation, read-back comparison, retrieval smoke testing, and failure escalation. Includes explicit untrusted-content/prompt-injection guardrails for third-party inputs.

92

1.19x

Quality

90%

Does it follow best practices?

Impact

97%

1.19x

Average score across 5 eval scenarios

Overview
Skills
Evals
Files

task.mdevals/scenario-4/

Knowledge Base Search Verification

Problem/Feature Description

A customer success team has integrated a knowledge base system to store notes from client interviews. The system ingests structured entries and is supposed to make them searchable so agents can pull up relevant context during follow-up calls. A recent deployment introduced a new storage backend, and there are reports from agents that searching for topics covered in recent interviews returns no results — even though the notes appear to have been saved.

You need to build a verification tool that saves a batch of interview notes to the knowledge base and then confirms each one can actually be found via a keyword search. The tool should report on whether each entry is discoverable, flag anything that isn't, and produce a summary the team lead can use to decide whether to roll back the deployment.

Output Specification

Produce two files:

  1. verify_kb.py — A Python script that writes each interview note entry below into a searchable store (you can implement search over an in-memory dict or list using substring/keyword matching), verifies each entry can be found by searching for a representative phrase from its content, and prints a JSON report to stdout.

  2. kb_verification_report.json — The report produced by running the script. It should include a per-entry result and an overall summary with a recommended action for the team.

The script must be runnable with python verify_kb.py using only Python standard library modules.

Input Files

The following interview notes should be stored and verified:

[
  {
    "key": "interview_cust_9201",
    "content": "Customer expressed concerns about onboarding timeline. Wants dedicated support during first 30 days. Follow up with CSM team.",
    "interviewer": "alex.chen",
    "date": "2026-03-04"
  },
  {
    "key": "interview_cust_9202",
    "content": "Positive feedback on dashboard redesign. Requested export to CSV feature for monthly reports. High renewal likelihood.",
    "interviewer": "priya.nair",
    "date": "2026-03-05"
  },
  {
    "key": "interview_cust_9203",
    "content": "Escalation risk: billing discrepancy from February invoice not yet resolved. Customer frustrated. Urgent ticket needed.",
    "interviewer": "alex.chen",
    "date": "2026-03-06"
  }
]

Install with Tessl CLI

npx tessl i markusdowne/memory-roundtrip-guard@0.1.2

evals

SKILL.md

tile.json