kairos
Copilot instructions1 starsChanged 7 months agoMITView on GitHub
You are working on Kairos, a persistent self-aware AI system. Read AGENTS.md before making changes. Key rules:
In this repositoryOne other file it gives its agents.
- AGENTS.mdAGENTS.md
The file
We split this copilot-instructions.md into the rules it gives an agent
12 rules. About 0 are common, 10 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈10
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 7
Never modify Layer 0 without an ADR and human approval (see docs/specs/layer-map.md)
- Distinctive: line 8
Every new public surface requires a test
- Distinctive: line 27
Do not use esbuild for NestJS integration tests — SWC is required for decorator metadata
- Distinctive: line 30
Never log credentials, even redacted hashes
- Distinctive: line 31
Never put resolved credentials in LLM context
- Distinctive: line 32
Test fixtures use clearly fake tokens to avoid Trivy false positives
- Very specific: line 33
All WebSocket events carry request_id for trace correlation
- Distinctive: line 36
Don't invent new error shapes
- Distinctive: line 37
Don't add abstraction layers not demanded by a concrete second use case
- Very specific: line 38
Don't change a spec without updating docs/specs/
- Distinctive: line 39
Don't commit generated files
- Distinctive: line 40
Don't use backticks in file references in markdown — use links
≈ “Common” is an estimate: at least 1 in 5 of the 3,827 Copilot instructions files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Copilot Instructions — Kairos
You are working on Kairos, a persistent self-aware AI system. Read AGENTS.md before making changes. Key rules:
Before Any Change
- Consult docs/architecture.md for normative behavior
- Never modify Layer 0 without an ADR and human approval (see docs/specs/layer-map.md)
- Every new public surface requires a test
Stack
- Control plane: NestJS 10 + TypeORM + pino + Passport/JWT + Zod
- Cognition: Python 3.12 + pydantic v2 + httpx + structlog +
uvpackage manager - Frontend: Next.js 14 (App Router) + Tailwind + shadcn/ui + socket.io-client
- Data: Postgres 16 + pgvector + Redis 7 + MinIO (S3-compatible)
- Executor: Alpine Docker image, non-root user 1000:1000
Code Style
- TypeScript:
exactOptionalPropertyTypes: true; use spread pattern for optionals; Zod for runtime validation - Python: Ruff + mypy strict; pydantic v2 models; async-first; structlog with request_id binding
- Error envelope:
{ code, message, details, request_id }— consistent across all services - Timestamps: ISO8601 UTC; IDs: UUID v4
Testing
- TS: vitest (unit) + supertest (integration) + Playwright (e2e)
- Python: pytest + pytest-asyncio
- Integration tests share one login in
beforeAllto avoid rate limit (5/15min) - Do not use esbuild for NestJS integration tests — SWC is required for decorator metadata
Security
- Never log credentials, even redacted hashes
- Never put resolved credentials in LLM context
- Test fixtures use clearly fake tokens to avoid Trivy false positives
- All WebSocket events carry
request_idfor trace correlation
Don't
- Don't invent new error shapes
- Don't add abstraction layers not demanded by a concrete second use case
- Don't change a spec without updating
docs/specs/ - Don't commit generated files
- Don't use backticks in file references in markdown — use links
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

