docs: sanitized public CLAUDE.md (working context moved to private bizdocs)

Author: vveerrgg

CLAUDE.md

@@ -1,71 +1,26 @@
-# social-alignment
+# CLAUDE.md — social-alignment
 
-A deterministic pre-action **five-lens compass** for AI agents. The
-alignment pillar of the NSE platform. Before an agent takes a significant
-action, five lenses evaluate it and the enclave decides whether to proceed,
-inform, ask, or halt.
+## What this is
+A deterministic five-lens pre-action compass for AI agents. Part of the open-source OpenClaw Nostr toolkit. MIT licensed.
 
-**Import:** `pip install social-alignment` → `from social_alignment import AlignmentEnclave`
-
-Zero runtime dependencies. Does **not** depend on `nostrkey`.
-
-## Build & Test
+## Install
+```
+pip install social-alignment
+```
 
-```bash
+## Develop
+```
+python -m venv .venv && . .venv/bin/activate
 pip install -e ".[dev]"
-ruff check .
-pytest -q
+python -m pytest -q     # tests
+python -m build         # sdist + wheel
 ```
 
-## Structure
-
+## Layout
 - `src/social_alignment/` — package source
-  - `types.py` — enums (`ActionDomain`, `Lens`, `Severity`, `EscalationLevel`) and frozen
-    dataclasses (`ActionContext`, `LensResult`, `Projection`, `EscalationDecision`,
-    `CheckResult`, `Decision`, `AlignmentConfig`)
-  - `enclave.py` — the five pure lens functions, `_aggregate`, `_escalate`, `evaluate`,
-    and `AlignmentEnclave` (main entry point + in-memory decision log)
-- `tests/` — pytest suite (known-answer + per-lens unit tests)
-- `examples/basic_usage.py` — runnable example
-- `clawhub/` — OpenClaw skill metadata
-
-## Publish
-
-```bash
-# PyPI (needs API token + OTP)
-python3 -m build
-python3 -m twine upload dist/social_alignment-X.Y.Z*
-
-# ClawHub
-npx clawhub publish ./clawhub --slug social-alignment --name "Social Alignment" \
-  --version X.Y.Z --tags latest --changelog "..."
-```
-
-**Version must be bumped in 4 places:** `pyproject.toml`, `src/social_alignment/__init__.py`
-(`__version__`), `clawhub/metadata.json`, and `clawhub/SKILL.md` (frontmatter `version:`).
-
-> Note: PyPI already carries `0.1.0` and `0.1.1`; ClawHub metadata had reserved `0.1.4`,
-> and two divergent `0.1.5` builds were published/pushed. This build is `0.1.6` to stay
-> ahead of all prior versions and requires a republish.
+- `tests/` — pytest suite
+- `clawhub/` — ClawHub skill metadata (SKILL.md, metadata.json)
 
 ## Conventions
-
-- Python 3.10+, hatchling build, ruff linter (100 char line length)
-- **Zero runtime dependencies.** Do not add any — this pillar must install standalone.
-- Import matches package name: `pip install social-alignment` → `from social_alignment import ...`
-- The five lenses are pure functions: same `ActionContext` + `AlignmentConfig` → same result.
-- Severity ordering (`IntEnum`): `CLEAR=0 < CAUTION=1 < YIELD=2 < STOP=3`. Aggregate = max.
-- `STOP` always defers to the human: `record_proceeded()` on a STOP without
-  `owner_overrode=True` raises `RuntimeError`. Enforced in code, no workaround.
-- `AlignmentEnclave.create(owner_npub=..., owner_name=..., **overrides)` filters unknown
-  overrides against `AlignmentConfig` fields — the orchestrator passes a superset and must
-  never trigger a `TypeError`. `check(...)` filters unknown context kwargs the same way.
-- Decision log is in-memory only (no persistence in this minimal build).
-- Frozen dataclasses everywhere in `types.py`.
-
-## Orchestrator contract
-
-`nse-orchestrator`'s `entity.py` calls
-`AlignmentEnclave.create(owner_npub=..., owner_name=...)` then
-`.check(domain=..., description=..., **context)`. The orchestrator's `alignment` extra already
-references `social-alignment`; no orchestrator code change is needed.
+- Python + pyproject (hatchling). Pure-Python crypto (`cryptography`), zero native build deps.
+- Public, MIT-licensed, open source.