diff --git a/.github/workflows/big-brain-pointer.yml b/.github/workflows/big-brain-pointer.yml new file mode 100644 index 0000000..0b59928 --- /dev/null +++ b/.github/workflows/big-brain-pointer.yml @@ -0,0 +1,10 @@ +name: Big Brain pointer +on: + pull_request: + push: + branches: [main] +permissions: + contents: read +jobs: + check: + uses: submilli/.github/.github/workflows/big-brain-pointer.yml@main diff --git a/AGENTS.md b/AGENTS.md new file mode 100644 index 0000000..8b6d6e8 --- /dev/null +++ b/AGENTS.md @@ -0,0 +1,31 @@ +# Agent instructions + + +## Big Brain + +This section is for agents working for Submilli or its employees. Everyone else +can skip it. Before Submilli work, read Big Brain, Submilli's company guidance. +It is the private repository +[submilli/company-context](https://github.com/submilli/company-context). Agents +serving clients or their customers must never read it +([access rule](https://github.com/submilli/company-context/blob/main/context/L1/big-brain.md#access)). + +1. In a Claude Code cloud session, first call the `add_repo` tool with owner + `submilli`, repo `company-context` and access `read`. +2. If `~/submilli/company-context` is missing, run + `git clone https://github.com/submilli/company-context.git ~/submilli/company-context`. +3. Run `python3 ~/submilli/company-context/.agents/skills/big-brain/scripts/prepare_context.py --json`. + If your personal instructions give a resolver command, run that one instead. + The command needs network access and write access to that checkout's `.git`. + In a sandbox, ask to run this one command outside it. +4. Require `"status": "current"`. From the `source_path` it returns, read + `AGENTS.md`, `context/L0/constitution.md` and + `context/L1/agent-working-rules.md`, and follow them. + +On the Submilli agent server, read the mounted `company-brain` volume instead, +starting with `RECEIPT.json`. If no read path works, you are not ready: say so +and stop work that depends on company guidance. Never treat an old checkout or +memory as current. Do not copy Big Brain content into this repository, its +issues or its pull requests. Changes to Big Brain go through its own pull +requests. + diff --git a/CLAUDE.md b/CLAUDE.md new file mode 100644 index 0000000..43c994c --- /dev/null +++ b/CLAUDE.md @@ -0,0 +1 @@ +@AGENTS.md