diff --git a/.gitmodules b/.gitmodules
index 9f0724cd8e..b937d9f4d5 100644
--- a/.gitmodules
+++ b/.gitmodules
@@ -1,3 +1,6 @@
 [submodule "evals/deep-swe"]
 	path = evals/deep-swe
 	url = https://github.com/datacurve-ai/deep-swe
+[submodule "evals/vendor/pier"]
+	path = evals/vendor/pier
+	url = https://github.com/lavaman131/pier
diff --git a/DEV_SETUP.md b/DEV_SETUP.md
index 3195307bd7..c9842ad580 100644
--- a/DEV_SETUP.md
+++ b/DEV_SETUP.md
@@ -7,6 +7,8 @@ This document covers setup, the local dev loop, testing patterns, and project la
 ## Prerequisites
 
 - **[Bun](https://bun.sh) ≥ 1.3.14** — runtime, package manager, and test runner
+- **[uv](https://docs.astral.sh/uv/)** — Python package/environment manager for the `evals/` harness
+- **Docker** — required for local Pier/DeepSWE sandbox runs
 
 This repo uses **Bun** for all development, scripts, and testing. The `@bastani/workflows` workspace package ships raw `.ts` files with no build step; Atomic bundles it into `@bastani/atomic` during the coding-agent build.
 
@@ -15,11 +17,53 @@ This repo uses **Bun** for all development, scripts, and testing. The `@bastani/
 ## Setup
 
 ```bash
-git clone git@github.com:bastani-inc/atomic.git
+git clone --recurse-submodules git@github.com:bastani-inc/atomic.git
 cd atomic
 bun install
 ```
 
+If you cloned without submodules, initialize them before running evals or touching vendored benchmark harnesses:
+
+```bash
+git submodule update --init --recursive
+```
+
+Current submodules include `evals/deep-swe` and `evals/vendor/pier`; the evals package points `datacurve-pier` at the local editable `evals/vendor/pier` checkout.
+
+The eval harness is Python-based and uses [`uv`](https://docs.astral.sh/uv/) from the `evals/` directory:
+
+```bash
+cd evals
+uv sync
+uv run python -c 'import pier, pathlib; print(pathlib.Path(pier.__file__).resolve())'
+uv run pier --help
+```
+
+The `pier` import should resolve to `evals/vendor/pier/src/pier/__init__.py`. After pulling submodule pointer changes or local Pier edits, refresh the editable install with:
+
+```bash
+cd evals
+uv sync --reinstall-package datacurve-pier
+```
+
+Example single-task DeepSWE run with Atomic and the local Pier checkout:
+
+```bash
+cd evals
+export OPENROUTER_API_KEY=...
+uv run pier run \
+  -p deep-swe/tasks \
+  --agent-import-path atomic_pier:Atomic \
+  --model openrouter/openai/gpt-5.5 \
+  --agent-kwarg thinking=xhigh \
+  --agent-kwarg version=next \
+  --agent-env 'OPENROUTER_API_KEY=${OPENROUTER_API_KEY}' \
+  --n-tasks 1 \
+  --sample-seed 0 \
+  --n-concurrent 1 \
+  --force-build
+```
+
 `bun install` runs the root `prepare` script, which installs Git hooks with [`prek`](https://prek.j178.dev/) from [`prek.toml`](./prek.toml). The hook shims installed by default come from `default_install_hook_types`; currently that is `pre-commit`. To reinstall hooks manually, run `bun run hooks:install`. Set `PREK_DISABLE_INSTALL=1` to skip hook installation for a local install; CI skips it automatically.
 
 The root `package.json` is a private workspace package named `atomic-monorepo`. The only publishable package is `packages/coding-agent` (`@bastani/atomic`); other `packages/*` workspaces are bundled or internal.
diff --git a/evals/pyproject.toml b/evals/pyproject.toml
index 81c1e1ab26..d592138705 100644
--- a/evals/pyproject.toml
+++ b/evals/pyproject.toml
@@ -9,4 +9,4 @@ dependencies = [
 ]
 
 [tool.uv.sources]
-datacurve-pier = { git = "https://github.com/datacurve-ai/pier" }
+datacurve-pier = { path = "vendor/pier", editable = true }
diff --git a/evals/uv.lock b/evals/uv.lock
index 1fba0afb2c..669e8ed2a1 100644
--- a/evals/uv.lock
+++ b/evals/uv.lock
@@ -378,7 +378,7 @@ wheels = [
 [[package]]
 name = "datacurve-pier"
 version = "0.3.0"
-source = { git = "https://github.com/datacurve-ai/pier#3f688a1bd0d15f77f7191609e9f8ecdd3a69db33" }
+source = { editable = "vendor/pier" }
 dependencies = [
     { name = "anthropic" },
     { name = "anyio" },
@@ -403,6 +403,38 @@ dependencies = [
     { name = "uvicorn" },
 ]
 
+[package.metadata]
+requires-dist = [
+    { name = "anthropic", specifier = ">=0.97.0" },
+    { name = "anyio", specifier = ">=4.13.0" },
+    { name = "botocore", specifier = ">=1.43.11" },
+    { name = "claude-agent-sdk", specifier = ">=0.1.17" },
+    { name = "daytona", specifier = ">=0.121.0" },
+    { name = "fastapi", specifier = ">=0.128.0" },
+    { name = "harbor", specifier = ">=0.5.0" },
+    { name = "jinja2", specifier = ">=3.1.0" },
+    { name = "litellm", specifier = ">=1.83.0" },
+    { name = "loguru", specifier = ">=0.7.3" },
+    { name = "mcp", specifier = ">=1.27.0" },
+    { name = "modal", specifier = ">=1.4.2" },
+    { name = "pydantic", specifier = ">=2.13.3" },
+    { name = "python-dotenv", specifier = ">=1.2.2" },
+    { name = "pyyaml", specifier = ">=6.0.3" },
+    { name = "rich", specifier = ">=14.3.4" },
+    { name = "shortuuid", specifier = ">=1.0.13" },
+    { name = "tenacity", specifier = ">=9.1.4" },
+    { name = "toml", specifier = ">=0.10.2" },
+    { name = "typer", specifier = ">=0.12" },
+    { name = "uvicorn", specifier = ">=0.38.0" },
+]
+
+[package.metadata.requires-dev]
+dev = [
+    { name = "pytest", specifier = ">=9.0.3" },
+    { name = "pytest-asyncio", specifier = ">=1.4.0" },
+    { name = "ruff", specifier = ">=0.5" },
+]
+
 [[package]]
 name = "datasets"
 version = "5.0.0"
@@ -593,7 +625,7 @@ dependencies = [
 ]
 
 [package.metadata]
-requires-dist = [{ name = "datacurve-pier", git = "https://github.com/datacurve-ai/pier" }]
+requires-dist = [{ name = "datacurve-pier", editable = "vendor/pier" }]
 
 [[package]]
 name = "fastapi"
diff --git a/evals/vendor/pier b/evals/vendor/pier
new file mode 160000
index 0000000000..bbc6857eaa
--- /dev/null
+++ b/evals/vendor/pier
@@ -0,0 +1 @@
+Subproject commit bbc6857eaab414f14633e1e9d697d1bf974266fc