mcp-ai-agent-guidelines 0.14.0 → 0.15.1

package/CHANGELOG.md

@@ -0,0 +1,140 @@
+# Changelog
+
+All notable changes to `mcp-ai-agent-guidelines` are documented here.
+This project follows [Keep a Changelog](https://keepachangelog.com/en/1.1.0/)
+and [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+---
+
+## [Unreleased]
+
+## [0.15.1] - 2026-04-19
+### Added
+- Added a dedicated Jest smoke-test lane for compiled ESM entrypoints alongside the main Vitest suite.
+- Expanded orchestration and workflow regression coverage around retry behavior, resolver selection, and infrastructure factories.
+- Lefthook integration upgraded to `repo-release-tools` v0.1.10: `rrt-update-unreleased` auto-writes changelog bullets on commit, `--strategy unreleased` pre-push guard replaces the per-commit file-diff check.
+
+### Fixed
+- enhance Jest testing support and improve orchestration config tests
+- Fixed MCP server direct-execution detection for symlinked bin invocations.
+- Clarified the published entrypoints: `mcp-ai-agent-guidelines` remains the MCP stdio server and `mcp-cli` is the interactive CLI.
+
+---
+
+## [0.15.0] — 2026-04-17
+
+### First pre-release of the MCP AI Agent Guidelines server.
+
+This release ships a fully functional TypeScript ESM MCP server exposing
+**102 AI agent skills** as callable tools across 18 domain prefixes, with a
+complete model-routing layer, governance surface, and observability pipeline.
+
+### Added
+
+- **102 skills** across 18 domain prefixes (`req-`, `orch-`, `doc-`, `qual-`,
+  `synth-`, `flow-`, `eval-`, `debug-`, `strat-`, `arch-`, `prompt-`,
+  `adapt-`, `bench-`, `lead-`, `resil-`, `gov-`, `qm-`, `gr-`)
+- **21 mission-driven instruction files** covering bootstrap, implement,
+  refactor, debug, test, design, review, research, orchestrate, adapt,
+  resilience, evaluate, prompt-engineering, plan, document, govern,
+  enterprise, physics-analysis, meta-routing, onboard_project, initial_instructions
+- **Model router** with role-name architecture (`free_primary`, `free_secondary`,
+  `cheap_primary`, `cheap_secondary`, `strong_primary`, `strong_secondary`,
+  `reviewer_primary`) backed by `orchestration.toml`
+- **orchestration-model-discover** MCP tool — self-reporting LLM discovery
+  that writes physical model IDs into `orchestration.toml` by role
+- **orchestration-config-read / orchestration-config-write** MCP tools for
+  non-interactive config management
+- **LlmProvider** type extended to cover `"xai"` and `"mistral"` in addition
+  to `"openai"`, `"anthropic"`, `"google"`, `"other"`
+- **`src/tools/model-discovery.ts`** — typed wiring for the model-discover tool
+- **`src/tests/globalSetup.ts`** — vitest global setup that restores
+  `orchestration.toml` from HEAD before each test run, preventing flakiness
+  from live MCP server tool invocations
+- **Observability pipeline** with structured JSON logging and
+  `ObservabilityOrchestrator`
+- **5 evaluation patterns** (parallel critique → synthesis, draft → review,
+  majority vote, cascade with fallback, free triple parallel + synthesis)
+- **PID homeostatic controller** for SLO-driven model tier selection
+- **Ant Colony Optimization router** (`adv-aco-router`) with pheromone decay
+- **Hebbian router** (`adv-hebbian-router`) for agent pairing reinforcement
+- **Physarum router** (`adv-physarum-router`) for self-pruning topology
+- **Redundant voter** (`adv-redundant-voter`) with NMR fault tolerance
+- **Membrane orchestrator** (`adv-membrane-orchestrator`) for data-boundary enforcement
+- **Replay consolidator** (`adv-replay-consolidator`) for orchestrator meta-learning
+- **Quorum coordinator** (`adv-quorum-coordinator`) for decentralised task assignment
+- **Clonal mutation** (`adv-clone-mutate`) for self-healing prompt recovery
+- **Simulated annealing optimizer** (`adv-annealing-optimizer`) for workflow topology search
+- **All 30 physics skills** (`qm-*`, `gr-*`) gated behind `physics-analysis`
+  instruction and guarded against default invocation
+- **`scripts/verify_matrix.py`** — zero-orphan skill coverage enforcement
+- **`scripts/generate-tool-definitions.py`** — auto-regenerates `src/generated/`
+  from `.github/skills/*/SKILL.md`
+- **Biome** for lint + format; **lefthook** for pre-commit quality gates
+- **Coverage reporting** with thresholds: statements 83 %, lines 84 %,
+  functions 87 %, branches 75 %
+- **`orchestration.toml`** as single orchestration authority with `strict_mode`,
+  capability tags, workload profiles, resilience config, and routing rules
+
+### Fixed
+
+- Opus 4.6 correctly classified as `strong_primary` (reasoning model),
+  not legacy `strong_secondary`
+- `orchestration.toml` role names standardised: `free_primary = gpt-5.1-mini`,
+  `strong_primary = sonnet-4.6` (`available = false` — reserved for synthesis
+  and adversarial review only)
+- CI workflow: `orchestration.toml` now tracked and validated via lefthook
+  `tracked-orchestration-config` pre-commit hook
+- All 1915 tests updated from physical model IDs to role-name IDs to match
+  the role-name `MODEL_PROFILES` architecture
+- `LlmLaneExecutor` switch branches and `ProviderEnvironmentVariable` added
+  for xAI and Mistral providers
+- Removed unused `canvas` devDependency that caused Docker builds to fail 
+  on  `node:24-alpine` due to missing native system libraries (cairo, pango, etc.)
+  required by node-gyp during `npm ci`.
+- Added targeted regression coverage for core runtime modules, including
+  `skill-cache`, `unified-orchestration`, and `llm-lane-executor`, to improve
+  branch coverage and guard against future regressions.
+- Added `.trivyignore` to exclude `CVE-2026-33671` from CI security scans until all dependencies are patched.
+
+### Changed
+
+- `MODEL_PROFILES` keyed by role names (`"free_primary"`, …) rather than
+  physical model strings — physical IDs live exclusively in `orchestration.toml`
+- `createBuiltinOrchestrationDefaults()` returns `models: {}` — physical
+  models must be discovered and written by the `orchestration-model-discover`
+  tool, not hardcoded in source
+
+### Test coverage
+
+- **1 915 tests** across **262 test files** — all passing
+- Global setup guard (`src/tests/globalSetup.ts`) prevents TOML state leaks
+  between interactive sessions and clean CI runs
+- `src/tests/fixtures/orchestration.toml` fixture added — mirrors production
+  `orchestration.toml` with `strict_mode = false` so test runs on CI and
+  fresh clones never throw the "strict mode forbids fallback" error
+- `tool-call-handler.test.ts` — `returns structured snapshot status payloads`
+  now mocks `loadFingerprintSnapshot` so the handler enters the `present`
+  branch (which includes `snapshotId`); previously failed in CI where no
+  snapshot file was on disk
+
+### CI
+
+- Added `Lint & Code Quality (Lefthook)` job to `ci.yml` that runs
+  `lefthook run pre-commit --command biome-check` — satisfies the required
+  GitHub status check gate before the Quality Gate job
+- `test` job bootstraps `orchestration.toml` from the test fixture before
+  running `vitest` to ensure a clean, hermetic environment on all matrix nodes
+- `quality` job now depends on the `lint` job for a serial fast-fail chain:
+  `lint → quality → test → drift`
+- CI test matrix updated from `[20, 22, 24]` to `[22, 24, latest]` — Node 20
+  reached end-of-life and is dropped; `latest` added to track the current
+  stable release automatically
+
+---
+
+## Links
+
+- Repository: <https://github.com/Anselmoo/mcp-ai-agent-guidelines>
+- Issues: <https://github.com/Anselmoo/mcp-ai-agent-guidelines/issues>
+- MCP specification: <https://modelcontextprotocol.io>

package/LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2025 Anselmoo
+Copyright (c) 2024 Anselmoo
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal