feat(backends): multi-backend intelligence layer — LLM + Agent backends, persona-loading agents, honest CLI commands
Add IntelligenceBackend abstraction with two categories: - LLMBackend (OllamaLocal, OllamaCloud): CI runs tool loop, provides tools, constructs prompts - AgentBackend (Opencode): agent runs own tool loop, CI serializes request Refactor all 18 agents from hardcoded stubs to persona loaders that delegate to the active backend or fail honestly when no backend is available. Refactor OrchestratorAgent.executeStage() from monolithic switch to agent delegation via STAGE_AGENT_MAP for intelligent stages (research, plan, execute, verify), with mechanical stages (specify, clarify, complete) staying inline. Wire CLI commands with --backend flag and auto-detection (opencode → ollama-local → ollama-cloud). Harden rollback/ship with real git operations. No command returns fake success.
This commit is contained in:
CI
@@ -15,7 +15,15 @@ CI (Continuous Intelligence) is a fully autonomous AI-driven software engineerin
|
||||
|
||||
```
|
||||
src/
|
||||
agents/ # 18 agent implementations (all extend BaseAgent)
|
||||
agents/ # 18 agent implementations (persona loaders delegating to backends)
|
||||
backends/ # Intelligence backend layer
|
||||
types.ts # IntelligenceBackend, BackendRequest, BackendResult, BackendConfigSection
|
||||
tool-registry.ts # CI-owned tool implementations (readFile, writeFile, editFile, runBash, glob, grep)
|
||||
ollama-base.ts # Abstract base for Ollama backends (shared tool loop, prompt construction)
|
||||
ollama-local.ts # OllamaLocalBackend (localhost:11434)
|
||||
ollama-cloud.ts # OllamaCloudBackend (remote endpoint, auth, rate limiting)
|
||||
opencode.ts # OpencodeBackend (shells out to opencode --non-interactive)
|
||||
index.ts # Backend registry + auto-detection
|
||||
cli/ # Commander.js CLI (commands.ts, index.ts)
|
||||
core/ # Core engine components
|
||||
artifacts.ts # Legacy .planning/ artifact management (retained for backward compat)
|
||||
@@ -32,7 +40,7 @@ src/
|
||||
git-context.ts # Project state reconstruction from git log + branches
|
||||
types/ # Type definitions
|
||||
commit-meta.ts # CiMetadata, CommitDecision, CommitEscalation, ParsedCiCommit
|
||||
config.ts # CIConfig, AutonomyLevel, ModelProfile, DEFAULT_CI_CONFIG
|
||||
config.ts # CIConfig, AutonomyLevel, ModelProfile, DEFAULT_CI_CONFIG (includes backend)
|
||||
decisions.ts # Decision, ConfidenceLevel, DecisionCategory
|
||||
escalation.ts # Escalation, EscalationType, EscalationResolution
|
||||
clarify.ts # ClarifyQuestion, ClarifyResult
|
||||
@@ -45,7 +53,7 @@ src/
|
||||
security.ts # Layer 3: STRIDE threat analysis (stub)
|
||||
quality.ts # Layer 4: multi-persona code review (stub)
|
||||
index.ts # Public API exports
|
||||
version.ts # VERSION = "0.2.0"
|
||||
version.ts # VERSION = "0.3.0"
|
||||
templates/ # Template files (config.json, DECISIONS.md, specification.md)
|
||||
```
|
||||
|
||||
@@ -62,7 +70,7 @@ templates/ # Template files (config.json, DECISIONS.md, specification.md
|
||||
|
||||
- **Language**: TypeScript with ES2022 target, Node16 modules
|
||||
- **Module resolution**: Node16 style with `.js` extensions in imports
|
||||
- **Agent pattern**: All agents extend `BaseAgent` with `name`, `description`, and `execute(context: AgentContext): Promise<AgentResult>`
|
||||
- **Agent pattern**: All agents extend `BaseAgent` with `name` (AgentName), `description`, `workflow`, and `execute(context: AgentContext): Promise<AgentResult>`. Agents delegate to `context.backend` when available, fail honestly when not.
|
||||
- **No runtime validation library**: Uses plain TypeScript types, not Zod schemas (Zod is a dependency but types are hand-defined)
|
||||
- **File I/O**: Use `src/utils/file.ts` helpers (`writeFile`, `readFile`, `ensureDir`, `readJSON`, `writeJSON`) instead of raw `fs` calls in agent/business logic
|
||||
- **Config**: `CIConfig` type and `DEFAULT_CI_CONFIG` in `src/types/config.ts` — always merge partial configs with defaults
|
||||
@@ -77,7 +85,26 @@ templates/ # Template files (config.json, DECISIONS.md, specification.md
|
||||
SPECIFY → CLARIFY → RESEARCH → PLAN → EXECUTE → VERIFY → COMPLETE
|
||||
```
|
||||
|
||||
Each stage is executed by `OrchestratorAgent.executeStage()`. The orchestrator iterates through `STAGE_ORDER` and collects `PhaseResult` for each.
|
||||
Each stage is executed by `OrchestratorAgent.executeStage()`. The orchestrator delegates intelligent stages (research, plan, execute, verify) to specialized agents via `context.backend` when available, falling back to mechanical execution when no backend is configured. Mechanical stages (specify, clarify, complete) are always handled by the orchestrator directly.
|
||||
|
||||
## Intelligence Backend Architecture
|
||||
|
||||
```
|
||||
IntelligenceBackend (unified interface)
|
||||
├── LLMBackend (CI runs tool loop, provides tools, constructs prompts)
|
||||
│ ├── OllamaLocalBackend (localhost:11434, no auth)
|
||||
│ ├── OllamaCloudBackend (remote endpoint, API key, rate limits)
|
||||
│ └── (future: OpenAI, Anthropic, Gemini, etc.)
|
||||
└── AgentBackend (agent runs own tool loop, CI sends request)
|
||||
├── OpencodeBackend (opencode --non-interactive)
|
||||
└── (future: Codex, Claude Code, Hermes, etc.)
|
||||
```
|
||||
|
||||
- **LLM backends**: CI constructs system prompts from persona.md + workflow.md, defines tool schemas, runs the tool-call loop via `ToolRegistry`, and parses structured JSON output
|
||||
- **Agent backends**: CI serializes `BackendRequest`, invokes the agent, and parses JSON `BackendResult` from stdout
|
||||
- **Auto-detection** (provider: "auto"): tries opencode → ollama-local → ollama-cloud → fails with instructions
|
||||
- **Per-command override**: `ci run --backend ollama-local` forces a specific backend
|
||||
- **Config**: `backend` section in `.ci/config.json` with provider, fallback, agent_backends, llm_backends
|
||||
|
||||
## Agent Modification Rules (from PRD)
|
||||
|
||||
@@ -174,5 +201,6 @@ Each stage is executed by `OrchestratorAgent.executeStage()`. The orchestrator i
|
||||
- **Reconstruction test**: An agent with only commit message access can reconstruct project state (phase, decisions, requirements coverage, lessons, escalations)
|
||||
- **Verification layers**: All 4 layers implemented — structural, behavioral, security (STRIDE), quality
|
||||
- **CLI**: All 11 commands wired up (`init`, `run`, `quick`, `debug`, `verify`, `review`, `status`, `audit`, `clarify`, `rollback`, `ship`)
|
||||
- **Agent implementations**: Stub agents return success immediately. Real LLM-based agent implementations are needed for research, planning, execution, verification, etc.
|
||||
- **Agent implementations**: Persona loaders that delegate to active backend. Fail honestly when no backend is available (no more fake success).
|
||||
- **Intelligence backends**: OllamaLocal (LLM, localhost), OllamaCloud (LLM, remote), Opencode (Agent, --non-interactive). Auto-detection: opencode → ollama-local → ollama-cloud.
|
||||
- **Tests**: 25 test suites, 218 tests covering types, config, decision-engine, escalation, clarify, commit-parser, commit-builder, git-context, git-branch, ci-files, all 4 verification layers, file utils
|
||||
Reference in New Issue
Block a user