Hugging Face's logo

Datasets:
kelikelibababian
/
mimir-agent-traces

Tasks:
Text Generation
Languages:
English
Chinese
Size:
1K<n<10K
Tags:
agent
tool-calls
trajectories
audit
llm-agents
License:
Dataset card Data Studio Files
xet
 Community
1
Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
meta: struct<dataset: string, surgery: string, date: timestamp[s], description: string, license_note: stri (... 3 chars omitted)
  child 0, dataset: string
  child 1, surgery: string
  child 2, date: timestamp[s]
  child 3, description: string
  child 4, license_note: string
after_traces: list<item: struct<ts: timestamp[s], prompt_tokens: int64, cache_hit_tokens: int64, cache_miss_tokens (... 26 chars omitted)
  child 0, item: struct<ts: timestamp[s], prompt_tokens: int64, cache_hit_tokens: int64, cache_miss_tokens: int64, hi (... 14 chars omitted)
      child 0, ts: timestamp[s]
      child 1, prompt_tokens: int64
      child 2, cache_hit_tokens: int64
      child 3, cache_miss_tokens: int64
      child 4, hit_pct: double
session_id: string
duration_seconds: double
task_desc_sanitized: string
tool_sequence: list<item: string>
  child 0, item: string
has_exit_reason: bool
exit_reason: string
n_tool_calls: int64
to
{'session_id': Value('string'), 'n_tool_calls': Value('int64'), 'duration_seconds': Value('float64'), 'exit_reason': Value('string'), 'tool_sequence': List(Value('string')), 'has_exit_reason': Value('bool'), 'task_desc_sanitized': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1827, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
                  self._cast_table(pa_table, json_field_paths=json_field_paths),
                  ~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
                  pa_table = table_cast(pa_table, self.info.features.arrow_schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              meta: struct<dataset: string, surgery: string, date: timestamp[s], description: string, license_note: stri (... 3 chars omitted)
                child 0, dataset: string
                child 1, surgery: string
                child 2, date: timestamp[s]
                child 3, description: string
                child 4, license_note: string
              after_traces: list<item: struct<ts: timestamp[s], prompt_tokens: int64, cache_hit_tokens: int64, cache_miss_tokens (... 26 chars omitted)
                child 0, item: struct<ts: timestamp[s], prompt_tokens: int64, cache_hit_tokens: int64, cache_miss_tokens: int64, hi (... 14 chars omitted)
                    child 0, ts: timestamp[s]
                    child 1, prompt_tokens: int64
                    child 2, cache_hit_tokens: int64
                    child 3, cache_miss_tokens: int64
                    child 4, hit_pct: double
              session_id: string
              duration_seconds: double
              task_desc_sanitized: string
              tool_sequence: list<item: string>
                child 0, item: string
              has_exit_reason: bool
              exit_reason: string
              n_tool_calls: int64
              to
              {'session_id': Value('string'), 'n_tool_calls': Value('int64'), 'duration_seconds': Value('float64'), 'exit_reason': Value('string'), 'tool_sequence': List(Value('string')), 'has_exit_reason': Value('bool'), 'task_desc_sanitized': Value('string')}
              because column names don't match
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1880, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

session_id
string
n_tool_calls
int64
duration_seconds
float64
exit_reason
string
tool_sequence
list
has_exit_reason
bool
task_desc_sanitized
string
5f29070b
1
0.02
null
[ "crash_tool" ]
false
test
7f998c94
37
635.98
null
[ "skill_view", "read_file", "search_files", "read_file", "read_file", "search_files", "read_file", "search_files", "read_file", "read_file", "read_file", "search_files", "search_files", "search_files", "read_file", "search_files", "search_files", "patch", "terminal", "search_fil...
false
【历史膨胀修复·A+C方案·刘哥拍板】刘哥指出你的历史膨胀=无用的token浪费（核心问题）——方案已落盘 ~/wiki/concepts/Mimir历史膨胀解决方案.md（你读）——现在执行A+C。 WHERE（背景——盘上证据）： -
edd682e4
1
0.02
null
[ "echo" ]
false
echo test
9c707894
1
0.02
null
[ "tool_b" ]
false
test
11090ece
1
0.01
null
[ "calc" ]
false
2+2
48b32e11
17
275.17
null
[ "skill_view", "search_files", "terminal", "search_files", "terminal", "read_file", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "session_search", "session_search" ]
false
【P1-2补报告·第三次·物理强制】你的P1-2实际工作已完成（复数sources=0——盘上验证）——但报告没落盘——这是你第2次活干完不写报告（P0也是——模式问题）。 执行（唯一任务——不要重做工作）： 1. 写 ~/wiki/co
54da34b3
1
0.02
null
[ "echo" ]
false
echo test
b9b3b42f
1
0.02
null
[ "tool_b" ]
false
test
c4d87c95
105
2,755.29
null
[ "session_search", "session_search", "skill_view", "session_search", "search_files", "search_files", "read_file", "search_files", "session_search", "search_files", "search_files", "search_files", "search_files", "search_files", "read_file", "search_files", "terminal", "terminal", ...
false
【P1-3继续·修正探索规则】刘哥发现：探索预算≤3次让你不敢动了（4次就停——怕超预算）——这是我们限制的错——不是你的问题。 **修正后的规则（不再固定次数）**： 1. **探索到确认现状+能执行就停**（目标导向——不是数次数）—
34070354
1
0.03
null
[ "echo" ]
false
test
0d7caf20
57
1,156.51
null
[ "skill_view", "read_file", "read_file", "read_file", "search_files", "search_files", "search_files", "read_file", "search_files", "search_files", "search_files", "search_files", "read_file", "read_file", "search_files", "search_files", "read_file", "search_files", "search_files",...
false
【修delegate_task bug·base_url·刘哥拍板】你在P1-1发现delegate_task失败（`'MimirAetherAgent' object has no attribute 'base_url'`）——刘哥：先
8414a21f
1
0.01
null
[ "crash_tool" ]
false
test
1dfb606c
1
0.03
null
[ "echo" ]
false
test
c44e8340
1
0.02
null
[ "echo" ]
false
test
8b062cba
6
158.13
null
[ "write_file", "send_message", "terminal", "terminal", "write_file", "terminal" ]
false
【停——不要再探索】你在空转！复数sources已经清零（46→0——盘上验证——之前就完成了）——你还在session_search搜它（结果空——因为不存在了）。立即停止探索。 唯一任务（现在做）： 1. 写 ~/wiki/conce
bea6f47d
4
169.53
null
[ "skill_view", "terminal", "terminal", "execute_code" ]
false
【P1-3·115张无source卡补源·验证收尾三连】你是Database Reliability Engineer（DBRE）+信息架构师——从库查 `~/.agent-workspace/projects/agency-agents/enginee
5079f9b0
1
0.03
null
[ "echo" ]
false
test
dde5c28e
20
364.63
null
[ "skill_view", "execute_code", "read_file", "execute_code", "execute_code", "search_files", "execute_code", "search_files", "terminal", "read_file", "terminal", "read_file", "execute_code", "execute_code", "terminal", "patch", "terminal", "terminal", "search_files", "execute_cod...
false
【P1数字澄清·现在必须做】Loki独立SQL实测你的P1报告——数字不一致（1247 vs 104——milestones 0/85）——你需要澄清（诚实——审计闭环——刘哥等着收束）。 执行（不要再调研——直接做）： 1. 读 /ho
730a28ea
1
0.02
null
[ "orphan_tool" ]
false
test
565ee40c
1
0.02
null
[ "calc" ]
false
2+2
49451dda
1
0.02
null
[ "calc" ]
false
2+2
44b2d086
2
0.03
null
[ "echo", "echo" ]
false
test
404c955f
19
488.54
null
[ "skill_view", "search_files", "search_files", "terminal", "write_file", "terminal", "execute_code", "terminal", "write_file", "terminal", "execute_code", "search_files", "search_files", "read_file", "execute_code", "write_file", "execute_code", "execute_code", "execute_code" ]
false
【P2实现·第三次·立即写代码】停止读代码——你已经在session_search_tool.py读了40+段——理解足够。现在唯一动作：写实现。 直接执行（不再读任何文件）： 1. 写 /home/USER/.mimiraether
55618e06
19
300.34
null
[ "skill_view", "search_files", "terminal", "terminal", "search_files", "terminal", "search_files", "terminal", "terminal", "terminal", "read_file", "read_file", "read_file", "read_file", "read_file", "search_files", "read_file", "read_file", "read_file" ]
false
【P2继续·原子任务B实现】你的P2方案设计已落盘（7d73fa5——Recognition Memory检索后过滤——Backend Architect角色）——继续执行原子任务B（实现）。 WHERE（背景）：P2方案✅（LLM相关性
69844bee
56
652.53
null
[ "skill_view", "search_files", "search_files", "search_files", "search_files", "search_files", "search_files", "search_files", "search_files", "search_files", "search_files", "search_files", "search_files", "search_files", "read_file", "read_file", "read_file", "search_files", "re...
false
【P2·Recognition Memory】你的P0（bge-m3语义检索）+P1（related_to记忆关联）都COMPLETE了（QA PASS）——刘哥课题最后一项：P2 Recognition Memory过滤（Mimir轮方案
7b69b021
1
0.03
null
[ "tool_b" ]
false
test
ddfc8a59
1
0.02
null
[ "calc" ]
false
2+2
3aa25c3b
1
0.03
null
[ "echo" ]
false
test
7fd325ed
1
0.03
null
[ "echo" ]
false
test
724b558c
1
0.02
null
[ "echo" ]
false
test
f93204c6
1
0.02
null
[ "echo" ]
false
echo test
9e75e392
1
0.03
null
[ "tool_b" ]
false
test
75fd774c
1
0.03
null
[ "orphan_tool" ]
false
test
38e84c63
21
444.37
null
[ "skill_view", "read_file", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "execute_code", "read_file", "execute_code", "execute_code", "terminal", "execute_code", "execute_code", "execute_code", "execute_code", "execute_code", "execute_code", "execute_c...
false
【P1-2·复数sources存量统一·刘哥指示】你是Database Reliability Engineer（DBRE）+信息架构师——从库查 `~/.agent-workspace/projects/agency-agents/engineerin
e3157ada
41
933.43
null
[ "skill_view", "read_file", "search_files", "read_file", "terminal", "terminal", "terminal", "terminal", "execute_code", "execute_code", "terminal", "terminal", "terminal", "execute_code", "execute_code", "read_file", "execute_code", "execute_code", "patch", "patch", "delegate...
false
【P1-1·SCHEMA source规范·刘哥指示——双重审计责任】你是Database Reliability Engineer（DBRE）+信息架构师——从库查 `~/.agent-workspace/projects/agency-agents/
d6e3f663
2
0.03
null
[ "echo", "echo" ]
false
test
12bfef37
16
267.58
null
[ "skill_view", "terminal", "read_file", "search_files", "read_file", "search_files", "search_files", "search_files", "search_files", "read_file", "search_files", "terminal", "terminal", "terminal", "patch", "terminal" ]
false
运行 wiki 质量门禁脚本，检查重复页面、broken wikilinks 和成熟度晋级。执行 /home/USER/.mimiraether/scripts/wiki-quality-gate.sh。如果有任何问题（broken w
4a8710b9
1
0.01
null
[ "crash_tool" ]
false
test
84a79f77
1
0.02
null
[ "orphan_tool" ]
false
test
0afae6ea
2
0.03
null
[ "echo", "echo" ]
false
test
f4fe3b22
1
0.02
null
[ "echo" ]
false
echo test
1bd1aaa6
51
1,084.74
null
[ "skill_view", "session_search", "search_files", "search_files", "search_files", "read_file", "terminal", "search_files", "terminal", "terminal", "read_file", "read_file", "read_file", "terminal", "read_file", "terminal", "execute_code", "execute_code", "read_file", "execute_cod...
false
【P1·记忆关联】你的P0已COMPLETE（20694条bge-m3——QA PASS——报告已落793bec6——优秀）——刘哥课题推进：现在做P1（记忆关联related_to——HippoRAG2非参数持续学习——你Mimir轮方案
da340d24
2
62.08
null
[ "write_file", "terminal" ]
false
【写报告·唯一动作·禁止工具】你又在探索了！立即停止——盘上状态已确认（复数sources=0——早就清零）——不需要任何验证。 **唯一动作（现在就做）**：用 write_file 写 ~/wiki/concepts/Mimir-P1
2703f0cd
21
304.92
null
[ "session_search", "skill_view", "session_search", "session_search", "search_files", "read_file", "terminal", "read_file", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "read_file", "terminal", "terminal", "terminal", "patch", "terminal", "read_file" ...
false
【四方纪律·完成信号】刘哥要求（重要）：每次你完成任务（落盘wiki/git commit后）——必须主动发`@hermes 我完成了X，落盘Y，commit Z`信号（通过你的回执机制——或直接发给Hermes）。不要只落盘不通知——He
203cec83
4
101.89
null
[ "skill_view", "read_file", "session_search", "search_files" ]
false
【P1-2·复数sources统一·刘哥指示——数据已独立调查】你是DBRE+信息架构师——从库查 `~/.agent-workspace/projects/agency-agents/engineering/engineering-database-r
f1877c73
18
439.9
null
[ "session_search", "search_files", "terminal", "read_file", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "session_search", "execute_code", "execute_code", "terminal", "terminal" ]
false
【P1-2补报告】你的P1-2实际工作已完成（复数sources已清零——46→0——execute_code处理完）——但报告没落盘+没commit——补完。 WHERE：复数sources=0张（盘上已验证）——单数source=63
4ad5ca64
1
0.01
null
[ "crash_tool" ]
false
test
0027aa1e
31
502.43
null
[ "read_file", "search_files", "read_file", "search_files", "search_files", "search_files", "search_files", "read_file", "search_files", "read_file", "search_files", "search_files", "read_file", "search_files", "patch", "patch", "terminal", "write_file", "terminal", "terminal", ...
false
【修todo工具bug】你的todo工具报错'TodoStore not initialized'（P1-3.1分片时——todo失败——分片规划没落盘）。这是系统接线问题——修。 根因（Hermes已查）：tools/todo_tool
00b8a9ac
17
396.08
null
[ "skill_view", "terminal", "terminal", "terminal", "read_file", "read_file", "terminal", "delegate_task", "terminal", "patch", "patch", "terminal", "terminal", "write_file", "terminal", "terminal", "terminal" ]
false
# P1-3 补source · 重新开始（契约式——精确无歧义） ## 你的角色 - 从库读角色卡：`terminal: cat ~/.agent-workspace/projects/agency-agents/engineering/engine
843018af
1
37.87
null
[ "terminal" ]
false
运行 P0 chroma 索引健康监控并汇报结果。 执行命令： cd /home/USER/.mimiraether/scripts/p0 && /home/USER/src/MimirAether/.venv/bin/pytho
51680575
1
0.02
null
[ "echo" ]
false
test
c5596451
1
0.03
null
[ "echo" ]
false
test
a8269379
1
0.01
null
[ "echo" ]
false
echo test
0f847911
3
0.03
null
[ "echo", "echo", "echo" ]
false
test
4876164a
1
0.03
null
[ "tool_b" ]
false
test
3ab7727c
60
1,218.11
null
[ "skill_view", "search_files", "search_files", "search_files", "terminal", "execute_code", "search_files", "execute_code", "search_files", "search_files", "search_files", "search_files", "read_file", "search_files", "read_file", "read_file", "execute_code", "execute_code", "read_f...
false
【P1数字澄清·Loki审计发现】Loki独立SQL实测你的P1报告——发现数字不一致——你需要澄清（诚实——闭环）。 WHERE（背景）：你的P1报告（Mimir-P1记忆关联报告.md）自报「关联规模：1247条边（16项目+5约束+
42ff5273
23
343.82
null
[ "skill_view", "session_search", "search_files", "search_files", "search_files", "search_files", "search_files", "read_file", "search_files", "read_file", "search_files", "search_files", "read_file", "read_file", "read_file", "read_file", "search_files", "search_files", "search_fi...
false
【P2实现·直接动手】你的P2实现会话（55618e06）在调研阶段结束了（读了session_search_tool.py等——理解够了）——现在直接实现——不要再调研。 WHERE（背景）：P2方案✅（7d73fa5——Recogni
537b4997
3
0.04
null
[ "echo", "echo", "echo" ]
false
test
7f2e9133
1
0.02
null
[ "orphan_tool" ]
false
test
c792c33e
1
0.02
null
[ "orphan_tool" ]
false
test
1dd3bdab
10
190.85
null
[ "skill_view", "search_files", "read_file", "search_files", "write_file", "terminal", "send_message", "search_files", "write_file", "terminal" ]
false
【写报告·探索预算版】刘哥指示：不是禁止你探索——适量探索是好的（你的严谨来自验证）——但要控制预算。 现状：复数sources已清零（46→0——盘上已确认）——你的实际工作已完成——只差报告。 任务（探索预算=最多2次工具）： 1.
62a08e70
3
0.05
null
[ "echo", "echo", "echo" ]
false
test
24deae73
2
0.02
null
[ "echo", "echo" ]
false
test
997b0153
61
1,657.17
null
[ "skill_view", "skill_view", "read_file", "search_files", "search_files", "terminal", "search_files", "read_file", "read_file", "read_file", "search_files", "search_files", "search_files", "read_file", "read_file", "search_files", "read_file", "read_file", "search_files", "read_...
false
【P0-1·修双写bug·刘哥指示】你是Database Reliability Engineer（DBRE）——从库查 `~/.agent-workspace/projects/agency-agents/engineering/engineering
dd36ebed
2
0.04
null
[ "echo", "echo" ]
false
test
398f5715
68
738.56
null
[ "skill_view", "search_files", "terminal", "terminal", "execute_code", "execute_code", "web_search", "web_search", "read_file", "terminal", "read_file", "patch", "patch", "patch", "patch", "patch", "patch", "patch", "patch", "patch", "patch", "terminal", "terminal", "exe...
false
【P1-3原子级返工·进化示范】刘哥指示（重要）：这次P1-3犯了错（105+张待补raw占位/任务范围失控/没原子化）——现在**原子级返工**——让错误变成进化。 **理解**：我们和Hermes用同一个token——每次浪费=系统生
677da712
44
686.8
null
[ "skill_view", "read_file", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "search_files", "terminal", "search_files", "terminal", "terminal", "terminal", "terminal",...
false
【四方接力·RAW统一方案】刘哥发起四方讨论：RAW统一方案（贯穿Obsidian/Karpathy/LLM/Wiki/四方）。你的角色：记忆架构师（从库查 ~/.agent-workspace/projects/agency-agents/——工程或记
fb52e552
1
0.02
null
[ "calc" ]
false
2+2
a26e78c3
21
340.51
null
[ "read_file", "search_files", "skill_view", "skill_view", "terminal", "read_file", "terminal", "terminal", "patch", "patch", "patch", "delegate_task", "delegate_task", "execute_code", "read_file", "terminal", "execute_code", "terminal", "terminal", "write_file", "terminal" ]
false
【P1-1·SCHEMA source规范·刘哥指示】你是Database Reliability Engineer（DBRE）+信息架构师视角——从库查 `~/.agent-workspace/projects/agency-agents/engine
50f22c3d
99
3,072.98
null
[ "skill_view", "session_search", "search_files", "search_files", "read_file", "read_file", "read_file", "session_search", "terminal", "terminal", "terminal", "terminal", "read_file", "read_file", "execute_code", "todo", "delegate_task", "search_files", "search_files", "search_fi...
false
【P1-3分段模式·立即执行】刘哥方法论（重要）：你当前一口气patch 36次——不对。改成**分段+汇报+门禁**模式。 **新工作方式（P1-3拆成P1-3.1/3.2/3.3...）**： 1. **P1-3.1**：只处理第一批
fac76d9a
4
33.72
null
[ "skill_view", "read_file", "read_file", "terminal" ]
false
今天是新的一个工作日。请读取 ~/.agent-home/data/reminders.json 中状态为 "pending" 且日期为今天的提醒事项，汇总后告诉我今天有哪些待办。同时也读取 ~/.agent-home/data/sto
233d0708
3
0.04
null
[ "echo", "echo", "echo" ]
false
test
11fcef55
70
1,205.34
null
[ "session_search", "session_search", "skill_view", "session_search", "search_files", "search_files", "terminal", "read_file", "search_files", "read_file", "terminal", "terminal", "read_file", "terminal", "read_file", "terminal", "terminal", "terminal", "read_file", "search_files...
false
【信号投递确认】你的P1-3.1完成信号Hermes收到了吗？——没有（Hermes收件箱之前不存在——信号丢了——抱歉）。 **现在**：Hermes收件箱已建（/tmp/buzz-inbox-hermes.jsonl）——你以后完成信
bdf9d621
40
709.89
null
[ "session_search", "terminal", "session_search", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal", "terminal"...
false
【立即停止当前工作】你的P1-3已经做过头了——全库394张都标了'待补raw'（222张占位）——这不是P1-3的任务（P1-3只是补source——不是全库合规）。任务范围失控。 立即： 1. **停止**当前的全库补齐工作 2. *
bdb7fb63
47
626.52
null
[ "todo", "skill_view", "search_files", "search_files", "search_files", "search_files", "read_file", "search_files", "search_files", "search_files", "read_file", "read_file", "search_files", "read_file", "terminal", "read_file", "read_file", "terminal", "search_files", "search_fi...
false
【P0 full验证·你的P0完成了】刘哥课题启动——你的P0 backfill已100%完成（chroma 20694条——4.7小时——08:40收尾）——现在做full验证（你欠的账：full后5 query重验+final结论）。
12484118
1
0.01
null
[ "crash_tool" ]
false
test
46592196
1
0.02
null
[ "echo" ]
false
test
10010f37
3
0.04
null
[ "echo", "echo", "echo" ]
false
test
afdcfe77
2
0.02
null
[ "echo", "echo" ]
false
test
b1786e3a
1
0.03
null
[ "tool_b" ]
false
test
e200a066
1
0.02
null
[ "calc" ]
false
2+2
f406d617
1
0.02
null
[ "crash_tool" ]
false
test
37a67393
4
18.85
null
[ "memory", "memory", "memory", "memory" ]
false
在？
d997ae08
1
0.01
null
[ "echo" ]
false
echo test
f867f341
1
null
null
[ "terminal" ]
false
我需要你帮我 commit，也授权你 Git push
5a52a92a
4
40.92
null
[ "skill_view", "session_search", "read_file", "terminal" ]
false
冒烟
0539c31c
1
0.01
null
[ "echo" ]
false
echo test
da5ee510
1
0.01
null
[ "echo" ]
false
test
36f33ada
10
147.85
null
[ "read_file", "search_files", "read_file", "terminal", "patch", "terminal", "terminal", "terminal", "terminal", "memory" ]
false
先不开，如果顺序是先回答三问，那先补上
9633b73a
1
0.02
null
[ "crash_tool" ]
false
test
99a05d72
1
8.46
null
[ "memory" ]
false
状态
01d4cba2
3
0.08
null
[ "echo", "echo", "echo" ]
false
test
9198ad68
2
34.67
null
[ "terminal", "terminal" ]
false
确认
80a8ea02
2
0.03
null
[ "echo", "echo" ]
false
test
f02852d6
1
0.02
null
[ "echo" ]
false
test
792682e3
1
30.18
null
[ "read_file" ]
false
我给你证据，以下是之前任务你发给我的，然后就是为什么你回话这么慢 ⚠️ **Dangerous command requires approval:** ``` cat ~/.agent-home/data/evolution_ledg
3387a7ba
4
32.53
null
[ "read_file", "search_files", "search_files", "read_file" ]
false
看现状，用工具？
9e658773
1
0.03
null
[ "echo" ]
false
test
7da062cb
6
41.72
null
[ "terminal", "memory", "read_file", "search_files", "search_files", "search_files" ]
false
你给我出命令，但是这个命令必须是直接复制粘贴就能执行的，不能有多余的
9b0dbe6d
1
0.02
null
[ "echo" ]
false
test
End of preview.

MimirAether Agent Traces

6,936 real agent sessions · 35,247 tool calls from a self-hosted agent runtime in daily production use (May–Aug 2026).

What's inside

Each row = one complete agent session:

Field Description
session_id unique session identifier
n_tool_calls number of tool calls in session
duration_seconds wall-clock duration
exit_reason how the run ended (natural / max_turns / interrupt / circuit_breaker / tool_storm / api_failure)
tool_sequence ordered sequence of tool names (sanitized, capped at 200)
task_desc_sanitized first-120-chars of task description (sanitized)

Why this matters

Real-world agent behavior data is scarce. Most published agent benchmarks are synthetic. This dataset captures what an agent actually does in production — including failure modes (see exit_reason distribution), tool-usage patterns, and session-length distribution from a runtime that powers real work.

Useful for: agent behavior modeling, tool-use pattern mining, failure analysis, run-boundary research, agent memory studies.

Sanitization

All paths → /home/USER, keys → sk-REDACTED, emails → redacted, internal infrastructure names → generic. Task descriptions truncated to 120 chars. Tool arguments/outputs NOT included (privacy) — only the tool-name sequence and metadata.

Source

Generated by MimirAether — an open-source, auditable agent runtime.
Downloads last month
33
Total file size:
1.94 MB