All products

Review DeepSeek Harness chats with AI

Turn supported local conversation history into statistical cards and a narrative report you can inspect before sharing

Review DeepSeek Harness chats with AI when you want an overview rather than reopening every session. The agenttime skill reads supported local logs, calculates activity measures and combines factual cards with a narrative in a single HTML report.

The narrative must stay within the evidence. Gaps in chat history do not establish what happened in your life, how you felt or why you were absent. DeepSeek Harness is the only supported source at this pin; discovering another agent’s files does not make them readable. This is neither personality assessment nor proof of productivity.

What it does

  • Create a card-based report from supported local conversations
  • See activity measures grounded in identifiable user messages
  • Review the covered period and active sessions
  • Open the resulting HTML report in a browser
  • Identify unreadable logs and incomplete source coverage
  • Keep calculated figures separate from narrative interpretation

Install and use

  1. Install the skill in a file/command-capable agent with Python3.9 or later and the zstandard log-reading library. Your own local DeepSeek Harness history must be available.
  2. For example: “Report on my authorised local chats without guessing what happened on days with no messages.”
  3. Respect unsupported-source warnings rather than substituting another agent’s data without agreement.
  4. Open the reported output path and check quotations, coverage and private details before sharing.

FAQ

Source

Made by hongshuxifan321. Original skill: agenttime, published on GitHub under the MIT licence.

View on GitHub