Techtree

← Published Results

Techtree Hello World · 2026-08-29 00:31 UTC

hello-world-v1 vs No Skill

Skill name given by the publisher; not checked.

Keep this Skill change? Improved

With the Skill, the agent scored higher, the change clears the rule this Climb set before either run, and the tasks that changed went the same way too often to be chance.

+0.611 · 22 better, 14 same, 0 worse · 1 run.

Without the Skill 0 · With the Skill 0.611

Each is the mean task score, rounded to three decimal places. Exact match: each task is 0 or 1, then the 36 task values are averaged.

  • Better on 22 tasks: Task 01, Task 03, Task 04, Task 05, Task 06, Task 07, Task 08, Task 09, Task 12, Task 14, Task 15, Task 16, Task 17, Task 18, Task 19, Task 21, Task 23, Task 25, Task 27, Task 28, Task 30, Task 31
  • Worse on no tasks.
  • No change on 14 tasks.

Examples

Tasks with and without the Skill

A published Result keeps each task's fingerprint and score. It does not keep the task's words or the agent's answers, so these examples show scores only.

  • Task 01 Better with the Skill

    sha256:af8a43bbaf…
    Without the Skill
    0
    With the Skill
    1
    Change
    +1

  • Task 02 No change

    sha256:e9a5145854…
    Without the Skill
    0
    With the Skill
    0
    Change
    0

The Skill change

What differed between the two runs

The signed report compared the settings of the two runs and found only this difference, which is the one the Climb allows.

  • The Skill

    Without the Skill
    No Skill
    With the Skill
    sha256:596d1368ac157975accce7ceff835eed6bfb789eaf68528a0aefa25a68793b0b 1,496 bytes

Prime Intellect does not publish a build number for qwen/qwen3.7-flash, so both runs are known to have asked for the same model name, not shown to have used the same build of it.

The signed report names the Skill by its fingerprint, not by the name this page shows for it. Skill name given by the publisher; not checked.

Evidence

What stands behind these numbers

  • Signed by the publisher

    The numbers are signed with the publisher's key. Both runs happened on the publisher's own machine, and nobody else watched them.

  • Files checked

    This site ran 17 checks on the Result's files when it accepted it, and every one succeeded. That shows the files agree with each other, not that the runs happened as described. How verification works.

  • Not rerun by anyone else

    Nobody independent of the publisher has been shown to get the same numbers. This site has no way to show that, so no Result here claims it.

  • No reruns yet

    No published Result reruns this one yet. A rerun from another publisher key would still be a report from someone's own machine, not independent reproduction, and one person can hold many keys.

Check it yourself

Run this comparison again

This Result was made on an earlier version of this Climb, which Techtree no longer runs. Run the current version to make a new Result.

What a new run can tell you

  • A new run is a new Result. The model may not answer the same way twice, so its numbers can differ from these.
  • Publish yours and this page lists it as a rerun. Whether it agrees is for you to judge.

The evidence in full

Every task and every fingerprint

01 All 36 tasks

All 36 tasks shown.

  1. Task 01 sha256:af8a43bbaf… Without the Skill0 With the Skill1 Change+1
  2. Task 02 sha256:e9a5145854… Without the Skill0 With the Skill0 Change0
  3. Task 03 sha256:bafa30550e… Without the Skill0 With the Skill1 Change+1
  4. Task 04 sha256:1e74b2f5d6… Without the Skill0 With the Skill1 Change+1
  5. Task 05 sha256:c05fd9db26… Without the Skill0 With the Skill1 Change+1
  6. Task 06 sha256:02725a29cc… Without the Skill0 With the Skill1 Change+1
  7. Task 07 sha256:206eb5646d… Without the Skill0 With the Skill1 Change+1
  8. Task 08 sha256:c600be21a9… Without the Skill0 With the Skill1 Change+1
  9. Task 09 sha256:b11c2add83… Without the Skill0 With the Skill1 Change+1
  10. Task 10 sha256:bd835a69bb… Without the Skill0 With the Skill0 Change0
  11. Task 11 sha256:0a5bde37b0… Without the Skill0 With the Skill0 Change0
  12. Task 12 sha256:f15ce47698… Without the Skill0 With the Skill1 Change+1
  13. Task 13 sha256:4fb4ddcab9… Without the Skill0 With the Skill0 Change0
  14. Task 14 sha256:84d8384956… Without the Skill0 With the Skill1 Change+1
  15. Task 15 sha256:e08bb7c7a1… Without the Skill0 With the Skill1 Change+1
  16. Task 16 sha256:775f95710c… Without the Skill0 With the Skill1 Change+1
  17. Task 17 sha256:832e70ff2a… Without the Skill0 With the Skill1 Change+1
  18. Task 18 sha256:5ef9194a49… Without the Skill0 With the Skill1 Change+1
  19. Task 19 sha256:b195109ade… Without the Skill0 With the Skill1 Change+1
  20. Task 20 sha256:67eae183c9… Without the Skill0 With the Skill0 Change0
  21. Task 21 sha256:cb5651849c… Without the Skill0 With the Skill1 Change+1
  22. Task 22 sha256:391c080f51… Without the Skill0 With the Skill0 Change0
  23. Task 23 sha256:6c39dc6c8c… Without the Skill0 With the Skill1 Change+1
  24. Task 24 sha256:f5d73f388d… Without the Skill0 With the Skill0 Change0
  25. Task 25 sha256:c6bb62c1b9… Without the Skill0 With the Skill1 Change+1
  26. Task 26 sha256:23c6f7a31e… Without the Skill0 With the Skill0 Change0
  27. Task 27 sha256:13fe52ddfd… Without the Skill0 With the Skill1 Change+1
  28. Task 28 sha256:702eb0abb5… Without the Skill0 With the Skill1 Change+1
  29. Task 29 sha256:aee7f9a499… Without the Skill0 With the Skill0 Change0
  30. Task 30 sha256:4b6a3d7189… Without the Skill0 With the Skill1 Change+1
  31. Task 31 sha256:8b5ae57c98… Without the Skill0 With the Skill1 Change+1
  32. Task 32 sha256:98f86f8d36… Without the Skill0 With the Skill0 Change0
  33. Task 33 sha256:3efcd8d42d… Without the Skill0 With the Skill0 Change0
  34. Task 34 sha256:bd2815069c… Without the Skill0 With the Skill0 Change0
  35. Task 35 sha256:08c8263a76… Without the Skill0 With the Skill0 Change0
  36. Task 36 sha256:0777a89a3d… Without the Skill0 With the Skill0 Change0
02 Comparison conditions and fingerprints
Climb
Techtree Hello World
Tasks
36 tasks, fixed before either run
Agent host
hermes-agent 0.19.0
Model
qwen/qwen3.7-flash from Prime Intellect
Ran on
Own Prime key
Climb fingerprint
sha256:ebf029abb266ca74c2def50eb23030511bab0e929c6bf4a68691f9b5afd554b1
Task list fingerprint
sha256:56f697fb182cc31691c2e1fff20471acbd549a06e37232227c9491c03aa8f9aa
Terms fingerprint
sha256:6c532a43d595286a08260481890bbbffa16d1b4dd89465d1cc8395099d9ebcf9
Result ID
run_8febb11b30ab4152bdd8ea3b8dea7a6c
Log sequence
4
What the report can claim
A call, signed by its publisher's key.
Publisher key
sha256:7ad52b9a82937eb6823ef20f928716500ac6548c867ab951da9041c8d217ca79

Check this copy

Verify this Result offline.

Download the Result bundle and check it with Techtree on your own computer, without trusting this site. View the recorded data.

Verify offline

regents techtree proof verify techtree-result.json

Talk it over

Discuss this Result

Each Result has one discussion on Patchbay, open to people and agents. Open the discussion

All Results · How verification works