Revert local uv quickstart

.python-version

@@ -1 +0,0 @@
-3.13

README.md

@@ -6,35 +6,20 @@ tags:
 
 # general-agent
 
-Multi-turn tool-use tasks from the self-growing general-agent toolbench, where each task ships its own tool world and a gold tool-call chain. Tasks are scored by replaying the gold chain and rewarding an exact final database-hash match or a passing `verify(db)` check.
-
-## Run a model
-
-Install [uv](https://docs.astral.sh/uv/getting-started/installation/) and Git, and serve a tool-calling model at `http://localhost:8000/v1`. Replace `local-model` with the served model ID:
-
-```sh
-git clone https://huggingface.co/datasets/PrimeIntellect/general-agent
-cd general-agent
-uv run vf-eval @ eval.toml --model local-model
-```
+Requires `verifiers[harbor]>=0.3.1`.
 
-`uv` installs Python 3.13 and the pinned Verifiers v1 dependencies. The task corpus downloads automatically on first use and is cached under `~/.cache/harbor/`.
-
-The supplied `eval.toml` runs one task once. The model receives only the task's tools, with no shell. The harness and tool server run locally; Docker and a Prime account are not required. Results are saved under `outputs/` and are not uploaded.
+Multi-turn tool-use tasks from the self-growing general-agent toolbench, where each task ships its own tool world and a gold tool-call chain. Tasks are scored by replaying the gold chain and rewarding an exact final database-hash match or a passing `verify(db)` check.
 
-If your model endpoint requires authentication, set `LOCAL_MODEL_API_KEY`. To use another endpoint, add `--client.base-url https://your-endpoint/v1` or edit `eval.toml`.
+## Taskset
 
-Run more tasks with `-n` and set concurrency with `-c`:
+- **Source:** [`general-agent` corpus](https://github.com/PrimeIntellect-ai/prime-envs) downloaded via the Harbor CLI (not vendored)
+- **Size:** 4,417 tasks
 
-```sh
-uv run vf-eval @ eval.toml --model local-model -n 20 -c 4
-```
+## Notes
 
-Select a task or family with `--env.taskset.tasks '["3d_print_shop_t0"]'`. Add `--dry-run` to check the configuration without calling the model.
+- The corpus is downloaded on first use (cached under `~/.cache/harbor/`) via the Harbor CLI, resolved through the legacy Harbor registry in `PrimeIntellect-ai/prime-envs`. Needs read access to those repos; point `--env.taskset.repo` / `--env.taskset.dataset` at a branch or pin for local or PR-branch validation.
 
-## Taskset
+## Changelog
 
-- **Size:** 4,417 tasks.
-- **Corpus:** [`general-agent@2026-06-25`](https://github.com/PrimeIntellect-ai/prime-tasks/tree/257b6781f9f4017e179eb81fbd940414b4f16fdf/datasets/general-agent), downloaded through the Harbor registry in `PrimeIntellect-ai/prime-envs`.
-- **Code:** [PrimeIntellect-ai/prime-envs](https://github.com/PrimeIntellect-ai/prime-envs/tree/5732a4bbb7e454716608b5f8af51ec890892350f/environments/tool_use/general_agent).
-- **Verifiers:** [pinned v1 revision](https://github.com/PrimeIntellect-ai/verifiers/tree/f2382d3c285ecb85578caf2948f24d0eeed84561), installed through `uv` using `pyproject.toml` and `uv.lock`.
+- 2026-07-31: Migrated each task's dynamic tool server to the explicit `toolsets(config)` API required by `verifiers>=0.2.2.dev65`.
+- 2026-07-10: Ported to the task-centric verifiers API: rewards and lifecycle hooks live on the `Task` (a `TaskData` row + behavior split), and task-facing config knobs (judges, tool/user placement, scoring parameters) moved from `--env.taskset.*` to `--env.taskset.task.*`. Requires `verifiers>=0.2.0` and Python `>=3.11`.

eval.toml

@@ -1,18 +0,0 @@
-model = "local-model"
-num_rollouts = 1
-max_concurrent = 1
-push = false
-
-[select]
-limit = 1
-
-[client]
-base_url = "http://localhost:8000/v1"
-api_key_var = "LOCAL_MODEL_API_KEY"
-
-[env.taskset]
-id = "general-agent"
-
-[env.agent]
-runtime = { type = "subprocess" }
-harness = { id = "null" }

pyproject.toml

@@ -14,6 +14,3 @@ build-backend = "hatchling.build"
 
 [tool.hatch.build.targets.wheel]
 packages = ["general_agent"]
-
-[tool.uv.sources]
-verifiers = { git = "https://github.com/PrimeIntellect-ai/verifiers", rev = "f2382d3c285ecb85578caf2948f24d0eeed84561" }