|
1 | | -# Agent Instructions (foxmemory-deploy) |
| 1 | +# AGENTS.md |
2 | 2 |
|
3 | | -## Priority |
4 | | -1. Keep deployment examples reproducible. |
5 | | -2. Keep one-node and split modes both functional. |
6 | | -3. Document all env vars and defaults. |
| 3 | +This is the **canonical automation guide** for this repository. |
7 | 4 |
|
8 | | -## Rules |
9 | | -- Never hardcode private credentials. |
10 | | -- Keep compose files minimal and explicit. |
11 | | -- Update README + runbook when topology changes. |
| 5 | +## Read order (required) |
| 6 | +1. Read this `AGENTS.md` first. |
| 7 | +2. Read `README.md` for product context. |
| 8 | +3. Read `docs/*` relevant to the files you are changing. |
| 9 | + |
| 10 | +All tool-specific instruction files (Claude/Codex/Copilot/Cursor) are thin pointers to this file. |
| 11 | + |
| 12 | +## Core rules |
| 13 | +- Keep changes small, testable, and documented. |
| 14 | +- Preserve API/deploy contracts unless intentionally versioned. |
| 15 | +- Update docs and tests whenever behavior changes. |
| 16 | +- Never commit secrets. |
| 17 | + |
| 18 | +## Repository-specific constraints |
| 19 | + |
| 20 | +### foxmemory-infer |
| 21 | +- Scope: inference only (stateless embeddings service). |
| 22 | +- Do **not** add persistence/storage logic here. |
| 23 | +- Keep `/health` and `/embed` stable. |
| 24 | + |
| 25 | +### foxmemory-store |
| 26 | +- Scope: memory write/search and persistence APIs. |
| 27 | +- Do **not** add embedding inference logic here. |
| 28 | +- Keep backend adapters replaceable. |
| 29 | + |
| 30 | +### foxmemory-deploy |
| 31 | +- Scope: reproducible deployment topology and runbooks. |
| 32 | +- Keep one-node and split-mode examples current. |
| 33 | +- Keep compose files explicit and copy/paste safe. |
| 34 | + |
| 35 | +## PR checklist |
| 36 | +- [ ] Scope boundaries respected |
| 37 | +- [ ] Tests updated/passing |
| 38 | +- [ ] Docs updated |
| 39 | +- [ ] No secrets or local artifacts committed |
0 commit comments