swmcc / swmcc/agentic-development

Per-soul memory and interactive roster consolidation

Open
#22 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement thrawn
Dominant language
Python
Stars
0
Forks
0
Avg merge
2m
Merged PRs (30d)
1

Description

Part of #16, deliberately last: the interactive roster is the least load-bearing part of the factory.

  • Per-soul memory at ~/.claude/agent-memory/<name>/ with lessons.md (corrections given, patterns learned) and journal.md (append-only track record); the soul instructs read memory first, append what you learned last
  • Corrections loop: when the human says "stop flagging that, it is intentional" the correction lands in that soul's lessons and never recurs
  • Consolidate the ten agents in ~/.claude/agents/ into four or five real characters sharing the souls roster: builder, skeptic (review and security merged), test obsessive, operator and librarian
  • Each remaining agent file is a thin wrapper per #19 so the same character exists in sessions and in the factory

The measure of success from the design note: a plain persona that remembers corrections becomes distinctive within weeks; a beautiful persona with no memory is a mask.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the design note and the related issues #16 and #19, then inspect ~/.claude/agents/. Define the per-soul structure under ~/.claude/agent-memory//, including lessons.md and journal.md, before consolidating the ten agents into the listed characters. Done means corrections persist per soul and the remaining agent files are thin wrappers sharing that roster.

Written by the indexing model from the issue text.

Assessment

Domain
ai, tooling
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.