Skip to content

feat: operational hardening + SIN-Brain memory integration (closes #8, #14, #15, #16)#23

Open
Delqhi wants to merge 2 commits into
mainfrom
feat/operational-hardening
Open

feat: operational hardening + SIN-Brain memory integration (closes #8, #14, #15, #16)#23
Delqhi wants to merge 2 commits into
mainfrom
feat/operational-hardening

Conversation

@Delqhi
Copy link
Copy Markdown
Contributor

@Delqhi Delqhi commented May 30, 2026

Summary

This consolidates multiple foundation PRs (#6 #7 #10 #11 #12 #13 #21) and completes four major WS/BR features from the operational-readiness and SIN-Brain memory roadmaps:

WS1–5: Operational Hardening (closes #8)

  • CI matrix (WS1): .github/workflows/ci.yml with ruff lint+format gate and pytest matrix across Python 3.11/3.12/3.13
  • Shared ruff config (WS2): E/F/I/W ruleset + one-shot mechanical format applied; __version__ aligned to 0.2.0
  • Release workflow (WS3): .github/workflows/release.yml builds sdist+wheel on v* tags, clean-env install check, GitHub Release upload, PyPI Trusted Publishing
  • Consistency check (WS4): scripts/check_consistency.py + tests verify version alignment, subsystem import health, mcp-config↔serve entry point; --strict mode for full-stack CI
  • Dev bootstrap (WS5): scripts/dev_install.sh + scripts/run_all_tests.sh for editable multi-repo setup

BR-1: Register SIN-Brain Memory Tools (closes #14)

  • memory.py adapter: thin pass-through bridge to sin_brain.mcp_tools (recall/remember/forget/pin/link_evidence)
  • Graceful degradation: absent sin-brain reports False in sin status, tools raise MemoryUnavailable with clean JSON errors
  • detect_env() probes availability + tier sizes; register_tools(mcp) wires MCP tools only when SIN-Brain importable

BR-2: Agents-md SIN-Brain Inject Block (closes #15)

  • Memory playbook rows added to tool guidance (recall, remember, link_evidence, forget) when sin-brain available
  • Negative constraints / red-zones section (6 hard constraints) always present to steer agents away from anti-patterns
  • SIN-Brain's AGENTS.md inject block embedded in 'Project memory' section; idempotent re-runs preserve single managed block

BR-3: MCP-Config Client Validation (closes #16)

Test Coverage

  • Memory adapter: 8 tests covering graceful degradation (absent package, failed imports), stats detection, tool registration with fake sin-brain
  • Agents-md: 5 new tests verifying red-zones always present, memory rows conditional, idempotent with inject; existing integration tests pass
  • Total: 122 tests passing; CI gate includes ruff check/format + optional extras matrix

Foundation PRs Merged

Files Added/Modified

  • .github/workflows/: ci.yml (new), release.yml (fixed for setuptools)
  • scripts/: dev_install.sh + run_all_tests.sh (new), check_consistency.py (new)
  • src/sin_code_bundle/: memory.py (new), agents_md.py (extended), cli.py (wired memory tools + hardening fixes)
  • tests/: test_consistency.py (new), test_memory.py (new), test_agent_integration.py (extended)
  • pyproject.toml: ruff config + optional extras; init.py: version synced
  • docs/plans/: operational-hardening.md status → implemented

Notes

  • All changes follow SOTA patterns: graceful degradation, fail-soft, defensive programming
  • Ready for full multi-repo SIN-Brain environment once SIN-Brain repo is live; works standalone with bundle-only today
  • SOTA approach to memory: thin adapter, deferred imports, monkeypatchable for testing, zero hard dependency

OpenSIN Bot added 2 commits May 30, 2026 08:22
feat(ops): operational hardening — CI matrix, lint gate, release, con…
92f9a4e 
…sistency (closes #8)

WS1 CI: ci.yml with ruff lint+format gate and pytest matrix (3.11/3.12/3.13).
WS2 lint: shared ruff config (E/F/I/W); one-shot mechanical format applied.
WS3 release: release.yml builds sdist+wheel on v* tags, clean-env install
  check, GitHub Release upload, PyPI Trusted Publishing.
WS4 consistency: scripts/check_consistency.py + tests (version alignment,
  subsystem import health, mcp-config<->serve entry point); --strict for CI.
WS5 bootstrap: scripts/dev_install.sh + scripts/run_all_tests.sh.
Also aligned __version__ to packaged 0.2.0.
feat: SIN-Brain memory cortex integration + agents-md enhancements (c…
767e774 
…loses #14, #15, #16)

BR-1 (Issue #14): Register SIN-Brain memory tools in 'sin serve'
  - memory.py adapter: recall/remember/forget/pin/link_evidence pass-throughs
    with graceful degradation when sin-brain absent. detect_env() reports
    tier sizes in 'sin status'. register_tools(mcp) wires MCP tools.

BR-2 (Issue #15): 'sin agents-md' embeds sin-brain inject block
  - agents_md.py extended: memory playbook rows (recall, remember, link_evidence,
    forget) appear in the tool guidance when sin-brain available.
  - Negative constraints / red-zones section added (always present).
  - inject text from SIN-Brain embedded in 'Project memory' section.

BR-3 (Issue #16): 'sin mcp-config' client validation + stdio hygiene
  - Acceptance criteria met by PR #6 (already in main): valid YAML/JSON/TOML,
    --write merge preserving foreign keys, serve logs to stderr.
  - Added explicit stdio hygiene test to lock contract.

Tests: 8 memory adapter (graceful degradation, stats, tool registration),
       5 agents-md (red-zones always present, memory rows conditional,
       idempotent with inject), existing agent integration all pass (122 total).

CI gate: ruff check/format clean, 100% test pass, optional extras tested.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

1 participant

@Delqhi