EnterpriseBench evolves CodeScaleBench (275 tasks, 6 harnesses, Harbor contract) into a benchmark focused on context retrieval quality — measuring how well agents find and understand the right code across distributed codebases. Sourcegraph MCP is showcased through honest comparison as a controlled independent variable.
EnterpriseBench/
├── README.md # Project overview and quickstart
├── CLAUDE.md # Agent navigation guide
├── .claude/commands/ # Slash command skills
│ ├── diverge.md # Multi-perspective research
│ ├── diverge-prototype.md # Divergent prototyping in worktrees
│ └── converge.md # Structured debate via Agent Teams
├── schemas/
│ └── task.schema.json # Task definition schema
├── benchmarks/ # Task definitions by suite (100 tasks)
│ ├── dependency_management/
│ ├── incident_response/
│ ├── platform_engineering/
│ ├── security_operations/
│ ├── customer_escalation/
│ ├── feature_delivery/
│ ├── technical_debt/
│ └── mined/ # Mining candidate lists and provenance
├── lib/
│ └── eb_verify/ # Centralized verification library
├── scripts/
│ ├── mining/ # Task sourcing from OSS history
│ ├── sandbox/ # Multi-repo sandbox management
│ └── orchestration/ # Session chaining, event replay
├── configs/ # Run configurations
├── results/ # Run results and sample outputs
│ └── sample_runs/ # Sample verification outputs by task type
└── docs/ # Design docs, technical reports
- One verifier library —
eb_verifyis installed in every sandbox container. No copies. - Real OSS only — every task uses real open-source repositories with genuine dependency chains.
- Checkpoint scoring — every task has 2-5 graduated checkpoints for partial credit.
- Artifact diversity — tasks produce role-appropriate outputs, not just
answer.json. - Multi-repo default — the sandbox supports 1-5 repos per task natively.
- Tool-independent ground truth — curator never uses Sourcegraph or any evaluated tool.
Three tiers replace the single-source curator (F1=0.70):
| Tier | Method | What It Covers | LLM? |
|---|---|---|---|
| 1 — Deterministic | AST parsing, import graphs, dependency manifests (go.mod, package.json) | Structural dependencies, mechanically verifiable | No |
| 2 — LLM Curator | Semantic relevance analysis with cross-backend validation | Config files, docs, cross-cutting concerns | Yes |
| 3 — Solve-verification | Different model attempts task using ONLY curated context | Context sufficiency confirmation | Yes |
QA overlay: Confidence-weighted scoring, adversarial audit on sample, mutation testing on verifiers.
- Tool-independent generation (never uses Sourcegraph during ground truth creation)
- Cross-backend validation: local-only AND separate search backend, agreement >80% F1 = high confidence
- Two-tier annotation: every file labeled "required" or "sufficient"
- Chunk-level annotations: line ranges per file, not just paths
- Confidence metadata: each file carries confidence score + source
| Layer | Mechanism | Impact |
|---|---|---|
| Required files | Deterministic + curator agreement | Missing = significant penalty (binary match) |
| Sufficient files | Curator-identified, lower confidence | Missing = small penalty (soft matching) |
| Chunk-level | Line ranges within files | Block-level precision/recall |
| Checkpoints | 2-5 graduated per task | Partial credit for intermediate progress |
| Stratum | Share | Description |
|---|---|---|
| calibration | 15% | Single-repo, small codebase. MCP advantage should be <0.05. Bias check. |
| large_single | 25% | Large single-repo tasks |
| dual_repo | 30% | Two connected repos |
| multi_repo | 20% | 3-5 repos with dependency chains |
| monorepo_cross_package | 10% | Cross-package within monorepo |
Every task can run in three tool-access modes:
| Mode | Description | Dockerfile |
|---|---|---|
| baseline | No MCP tools, local search only | Standard |
| mcp_only | Mandatory Sourcegraph MCP | SG-Only |
| hybrid | Agent chooses tools freely | Hybrid (realistic) |
- 178 sg-evals mirrors from CSB, extended for multi-repo tasks
- Rich metadata captured per run: MCP tools called, frequency, latency, token cost
- 15% calibration tasks verify MCP bias < 0.05 on easy tasks
- Cross-stage integration tests: Canary tasks run end-to-end on every change
- Expanded fixture matrix: 6 artifact types validated
- Mutation testing: Verifier scripts tested with known-good and known-bad submissions
- Pre-flight checks: Sandbox health, repo presence, clone integrity
- Golden validation: Reference solutions for regression detection
Agent completes task
→ Sandbox health check (all repos present, no clone failures)
→ For each checkpoint (ordered):
→ Run checkpoint verifier script
→ Collect pass/fail + score
→ For each required artifact:
→ Validate artifact type (syntax, linting, structure)
→ Compute weighted total score
→ Write reward.txt
| Type | Container Lifecycle | State Mechanism | Scoring |
|---|---|---|---|
| single | One container, one run | N/A | Checkpoints at end |
| chain | N containers, sequential | Git branch between sessions | Milestones between + checkpoints at end |
| event_replay | One container, event stream | events.jsonl → actions.jsonl | Action correctness + timeliness |
| resume | One container, pre-populated branch | Git branch + progress doc | Checkpoints at end (same as single) |
EnterpriseBench inherits and extends CSB infrastructure:
- 275 tasks (220 Org + 55 SDLC) carry forward
- 6 existing harnesses remain compatible
- Harbor contract for container orchestration
- Task taxonomy migrated from SDLC/Org splits to 7 enterprise workflow clusters
- Metadata consolidated from 8+ scattered files to single source of truth