Datasets:
schema_version stringclasses 1
value | uuid stringlengths 42 42 | task_name stringclasses 630
values | tool_name stringclasses 2
values | responses_create_params dict | expected_answer stringlengths 181 6.32k | agent_ref dict | metadata dict |
|---|---|---|---|---|---|---|---|
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_c0378f5f3c4c05a0868c9831e5698d2f | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "I'm at the project root directory. I need to explore the project structure, understand the existing code, and then implement the missing pipeline logic.",
"plan": "First, I'll explore the directory structure to understand what files exist, then read the key files to understand the current implementat... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 0,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_22d7dbbdf741367b8e8da0a1c7c49101 | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "I can see the project structure. There are several key files: config/security.json, cli.py, normalizer.py, plugins/auditor.py, plugins/validators.py, plugins/minimizer.py, run.py, and tests/test_smoke.py. I need to read all of these to understand the existing code and what needs to be implemented.",
... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 1,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_87567f59afe083f9494c32ca44aafa7c | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "I've read the config file (contains crypto iterations and gadget memory boundaries), the empty __init__.py, and the stub files for cli.py and normalizer.py. I need to read the plugins and test files next.",
"plan": "Continue reading the remaining source files - the plugins and the test suite - to und... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 2,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_0204964db0ec0910b71c2a20ca832738 | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "I've read all the stub files. Now I need to read the test suite to understand the expected behavior and then look at some sample envelope data.",
"plan": "Read the test file and some sample envelope data to understand the data format and validation requirements.",
"commands": [
{
"keystro... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 3,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_ac70b3bdbfea1f06d06aee0672104a15 | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "I had a typo in my previous response. I need to explore the sample envelope data to understand the format before implementing the modules.",
"plan": "Look at sample envelopes to understand the data format, then implement all the modules.",
"commands": [
{
"keystrokes": "ls /data/telemetry... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 5,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_3772caf26ed54c0c680cf08444924ce2 | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "I can see the envelope format now. v1 uses 'envelope_ver', 'payload.telemetry.heap_trace/rop_chain/unwind_info', 'payload.transparency_log.leaf/siblings/root'. v2 uses 'schema', 'data.heap/rop/cfi', 'proof.target/path/expected'. I need to look at more envelopes to understand what makes some invalid.",
... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 6,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_6291751161d94ccbc5112a95a5a35697 | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "I now understand the data format for both v1 and v2 envelopes. I need to understand what makes some envelopes invalid. Let me look at a few more and also check if there are any corrupted files.",
"plan": "Check for corrupted/invalid envelopes and understand the patterns, then implement all modules.",... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 7,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_a3088c30495d0abfa39d166d67e015cb | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "All 950 UAF traces are minimizable, but 849 pass all validations. I need 800. Maybe the CFI validation is more complex - it might check that the stack pointer never goes negative during the unwind simulation, not just that the final sum is 0.",
"plan": "Check CFI validation with the constraint that s... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 12,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_d22c073a1f625cbdb3909193e8ba5cf9 | telemetry-pipeline-audit-2 | bash_command | {
"input": [
{
"role": "user",
"content": "You are an AI assistant tasked with solving command-line tasks in a Linux environment. You will be given a task description and the output from previously executed commands. Your goal is to solve the task by providing batches of shell commands.\n\nFormat your... | {
"analysis": "799 pass all checks but we need 800. Let me check the 50 no-UAF envelopes more carefully - maybe some have a different UAF pattern I'm missing, or maybe I need to reconsider the minimization logic.",
"plan": "Look at the no-UAF envelopes and also check if there are envelopes with UAF but failing some... | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {
"harness": "terminus_2",
"teacher_model": "zai-org/GLM-5.1",
"source_trajectory_uid": "f29960abccd3e479",
"pivot_agent_turn_index": 15,
"total_source_agent_turns": 41
} |
terminus2_pre_rp_pivot_sample_v1 | t2_pre_rp_4aa3cb848554bab87c81957dce823f8c | telemetry-pipeline-audit-2 | bash_command | {"input":[{"role":"user","content":"You are an AI assistant tasked with solving command-line tasks i(...TRUNCATED) | "{\n \"analysis\": \"env_15 has an empty heap_trace. Let me check all 50 no-UAF envelopes to unders(...TRUNCATED) | {
"type": "responses_api_agents",
"name": "terminus_judge_string_only_simple_agent"
} | {"harness":"terminus_2","teacher_model":"zai-org/GLM-5.1","source_trajectory_uid":"f29960abccd3e479"(...TRUNCATED) |
Dataset Description
The Nemotron-RL-Agentic-Terminal-Pivot-v1 dataset provides training samples for reinforcement learning of command-line ("terminal use") LLM agents with the terminus_judge environment in NeMo Gym.
Each record is a single agent decision point extracted from a successful agent trajectory on a terminal task:
responses_create_params.input— the prompt: the task instruction plus the terminal interaction history (prior agent actions and terminal outputs) up to the decision point;expected_answer— the reference next action taken by the teacher agent at that point (a JSON object in Terminus-2 format withanalysis,plan,commands[].keystrokes, andtask_completefields);agent_ref— routing to the NeMo Gymterminus_judgeresources server, which scores a policy model's generated action against the reference action to produce a reward for Reinforcement Learning from Verifiable Reward (RLVR).
The samples are drawn from ATCB, a collection of terminal-bench-style tasks; the dataset covers 630 unique ATCB seed tasks. Each seed task is a containerized Linux environment with a natural-language instruction, a hidden reference solution, and an automated verifier. The tasks emphasize realistic, long-horizon operational work in the terminal: building, repairing, and reconciling data pipelines; auditing and recovering from injected faults in logs, ledgers, caches, and telemetry stores; diagnosing crashed services, schedulers, and CI builds; and security/compliance workflows (access-log audits, SIEM triage, PII handling, TLS and authentication fixes) — set in scenarios spanning industrial and embedded systems (SCADA, firmware/OTA, cold-chain, satellite), HPC clusters, healthcare, finance, and media archives.
Trajectories were generated by running the seed tasks with the Terminus-2 agent (v2.0.0) in the Harbor execution harness, with GLM-5.1 (zai-org/GLM-5.1) as the teacher model. Only trajectories that passed the task's automated verifier were kept (at most 5 successful trajectories per task); each valid assistant turn of a kept trajectory was then converted into one next-action training sample. Degenerate turns (over-long command strings, duplicate prompts, post-completion confirmations) and records whose observation text captured collection-infrastructure details were filtered out.
All 630 seed tasks are ATCB tasks, and every released sample derives exclusively from ATCB trajectories. Internal collection metadata (provenance identifiers, pipeline and infrastructure fields) was likewise removed from every record; each record retains only the minimal fields documented under Dataset Quantification.
This dataset was utilized in the development of the NVIDIA Nemotron family of models: the same 630-task ATCB dataset was used for the reinforcement-learning post-training of both NVIDIA Nemotron Ultra and NVIDIA Nemotron Nano 3.5.
This dataset is ready for commercial or non-commercial uses.
Dataset Owner(s)
NVIDIA Corporation
Dataset Creation Date
Trajectories collected and samples constructed: 05/2026. Release preparation: 07/21/2026.
Versioning
v1.0
Previous Version(s): no previous version.
Relationship to Previous Version(s): Not applicable.
License/Terms of Use
This dataset is licensed under Creative Commons Attribution 4.0 International (CC-BY-4.0).
Intended Usage
Post-training large language models for agentic terminal use with reinforcement learning. The dataset is designed to be consumed by the NeMo Gym terminus_judge resources server (https://github.com/NVIDIA-NeMo/Gym/tree/main/resources_servers/terminus_judge): during training, a policy model is prompted with responses_create_params.input, and the judge scores the generated next action against expected_answer to produce a reward. The samples can also be used for supervised fine-tuning of terminal agents or for offline analysis of terminal-agent behavior.
Dataset Characterization
Data Collection Method
- Hybrid: Manually-Collected, Synthetic, Automated — seed tasks (instructions, environments, verifiers) in the ATCB collection were authored by a data vendor (Human); agent trajectories were generated by the GLM-5.1 teacher model driving the Terminus-2 agent in containerized task environments (Synthetic); execution, verification, and per-turn sample extraction were performed by automated pipelines (Automated).
Labeling Method
- Hybrid: Synthetic, Automated — the reference next action in each sample is teacher-model output (Synthetic); trajectory-level success labels come from each task's automated verifier script, and only verifier-passing trajectories were retained (Automated). No human labeling and no filtering model were used.
Dataset Format
Modality: Text. Format: JSON Lines (.jsonl), one training sample per line, UTF-8. Records are compatible with the NeMo Gym terminus_judge environment (https://github.com/NVIDIA-NeMo/Gym).
Dataset Quantification
Record count: 31,111 training samples (630 unique seed tasks; 2,716 distinct source trajectories; median 45 samples per task).
Feature count: 8 top-level fields — 7 data fields (including responses_create_params, expected_answer, agent_ref) plus a metadata object with 5 sub-fields.
Average prompt length ≈ 39,900 characters; average reference action ≈ 970 characters.
Total data storage: ~1.4 GB (JSONL, uncompressed).
Reference(s)
- NeMo Gym: https://github.com/NVIDIA-NeMo/Gym
- terminus_judge resources server: https://github.com/NVIDIA-NeMo/Gym/tree/main/resources_servers/terminus_judge
- Terminal-Bench: https://github.com/laude-institute/terminal-bench
- Terminus-2 agent / Harbor harness: https://github.com/laude-institute/harbor
- GLM-5.1: https://huggingface.co/zai-org/GLM-5.1
Ethical Considerations
NVIDIA believes Trustworthy AI is a shared responsibility and we have established policies and practices to enable development for a wide array of AI applications. Developers should work with their internal developer teams to ensure this dataset meets requirements for the relevant industry and use case and addresses unforeseen product misuse.
Task environments are synthetic, containerized scenarios authored by a data vendor; personal-appearing data in task content (names, e-mail addresses, credentials) was authored as fictional scenario material. Released text was additionally scanned for real personal and infrastructure information: no real personal data was identified, internal-only collection metadata (provenance identifiers, pipeline and infrastructure fields) was removed from all records, and records whose observation text captured collection-infrastructure details were dropped prior to release.
Please report quality, risk, security vulnerabilities or NVIDIA AI Concerns here.
- Downloads last month
- -