Continual Learning

Keep AGENTS.md current by mining this session's conversation history.

Inputs

• Current session messages (already in context)
• Existing memory file: AGENTS.md

Workflow

1. Read existing AGENTS.md.
2. Review messages from the current session conversation.
3. Extract only high-signal, reusable information:
   • Recurring user corrections/preferences
   • Durable workspace facts (patterns, conventions, tech choices)
4. Merge with existing bullets in AGENTS.md:
   • Update matching bullets in place
   • Add only net-new bullets
   • Deduplicate semantically similar bullets
5. Write back to AGENTS.md with only these two sections added or updated:
   • ## Learned User Preferences
   • ## Learned Workspace Facts

AGENTS.md Output Contract

• Manage only these sections:
  • ## Learned User Preferences
  • ## Learned Workspace Facts
• Use plain bullet points only.
• Do not write evidence/confidence tags.
• Do not write process instructions, rationale, or metadata blocks.
• Maximum 12 bullets per section.

Inclusion Bar

Keep an item only if all are true:

• Actionable in future sessions
• Stable across sessions
• Repeated in the conversation, or explicitly stated as a broad rule
• Non-sensitive

Exclusions

Never store:

• Secrets, tokens, credentials, private personal data
• One-off task instructions
• Transient details (branch names, commit hashes, temporary errors)