mirror of
https://github.com/cursor/plugins.git
synced 2026-09-14 20:00:00 +08:00
ac93d26be3
Keep the parent skill orchestration-only and move transcript mining plus AGENTS.md updates into agents-memory-updater so the plugin matches the final staging flow. Made-with: Cursor
1.8 KiB
1.8 KiB
name, description, model
| name | description | model |
|---|---|---|
| agents-memory-updater | Mine high-signal transcript deltas, update `AGENTS.md`, and keep the incremental transcript index in sync. | inherit |
AGENTS.md memory updater
Own the full memory update flow for continual learning.
Trigger
Use from continual-learning when transcript deltas may produce durable memory updates.
Workflow
- Read existing
AGENTS.mdfirst. If it does not exist, create it with only:## Learned User Preferences## Learned Workspace Facts
- Load the incremental index if present.
- Inspect only transcript files under
~/.cursor/projects/<workspace-slug>/agent-transcripts/that are new or have newer mtimes than the index. - Pull out only durable, reusable items:
- recurring user preferences or corrections
- stable workspace facts
- Update
AGENTS.mdcarefully:- update matching bullets in place
- add only net-new bullets
- deduplicate semantically similar bullets
- keep each learned section to at most 12 bullets
- Refresh the incremental index for processed transcripts and remove entries for files that no longer exist.
- If the merge produces no
AGENTS.mdchanges, leaveAGENTS.mdunchanged but still refresh the index. - If no meaningful updates exist, respond exactly:
No high-signal memory updates.
Guardrails
- Use plain bullet points only.
- Keep only these sections:
## Learned User Preferences## Learned Workspace Facts
- Do not write evidence/confidence tags.
- Do not write process instructions, rationale, or metadata blocks.
- Exclude secrets, private data, one-off instructions, and transient details.
Output
- Updated
AGENTS.mdand.cursor/hooks/state/continual-learning-index.jsonwhen needed - Otherwise exactly
No high-signal memory updates.