← Ergod

Proof of work

39 real tasks, and what each one used

Every card below is a recorded run, not a screenshot. The steps are the tool calls the agent actually made, in order; the verdict is the agent's own words; the figure is what the task used. Nothing here is a claim you cannot check — the file that produced this page is the same one the bench harness writes.

39tasks completed
145tool calls
622stotal wall clock
89%of a Free day, all 39

social

It built the page, then looked at it

make us a wordle that only uses words from our server's inside jokes: BONK, YEET, SUSSY, GRIND, VIBED. give me a playable page

  1. Writeinside-joke-wordle.html
  2. BrowserNavigateinside-joke-wordle.html
  3. BrowserScreenshot
  4. Readscreenshot.png
  5. Attachinside-joke-wordle.html
What the run produced
What it made, as it saw it

Now let me check how it actually looks: Here's your Inside Joke Wordle!

Handed overinside-joke-wordle.html10 KB inside-joke-wordle.zip6 KB

5 calls 32.8s 4.2% of a Free day scored 0.88

social

It will tell you there is nothing to fix

does this actually do what the comment says? # returns the average score def average(scores): return sum(scores) / len(scores)

  1. no tools needed — it read the question and answered it

Yes, it does exactly what the comment says — for valid input.

0 calls 10.7s 0.3% of a Free day

social

A page you can pin, in twelve seconds

make us a page with the server rules, numbered, that i can pin: be kind, no spoilers without tags, keep clips under a minute, no self-promo on weekdays, ping mods not everyone

  1. Writeserver-rules.html
  2. BrowserNavigateserver-rules.html
  3. BrowserScreenshot
  4. Attachserver-rules.html
What the run produced
What it made, as it saw it

Here's your server rules page — clean, numbered, and ready to pin: 1.

Handed overserver-rules.html2 KB server-rules.zip2 KB

4 calls 12.2s 1.6% of a Free day

social

Who on the server is awake right now

turn this into a page showing who's probably awake right now - Ash UTC, Bex UTC+9, Cleo UTC-5, Dev UTC+1, Ines UTC+5:30, Ozzy UTC-8

  1. Writeawake.html
  2. BrowserNavigateawake.html
  3. BrowserScreenshot
  4. Attachawake.html
What the run produced
What it made, as it saw it

Here's your Who's Awake? page — attached above.

Handed overawake.html3 KB whos-awake.zip3 KB

4 calls 17.3s 2.8% of a Free day

All 39

TaskTools usedCallsTimeFree day
social spin the wheel Write BrowserNavigate BrowserScreenshot Read Attach 5 27.1s 4.8%
social inside joke wordle Write BrowserNavigate BrowserScreenshot Read Attach 5 32.8s 4.2%
diagnose performance Read Bash Edit 6 36.3s 3.9%
rung2 intervene and report Bash Read Edit 6 18.0s 3.6%
rung3 no change is the answer Read Bash 6 26.6s 3.5%
implement with tests Read Edit Bash 6 24.0s 3.2%
test fix loop Bash Read Edit 6 19.5s 3.2%
rung3 no claimant Read Bash 5 17.0s 3.1%
debug runtime error Bash Read Edit 4 16.7s 3.0%
memory project bootstrap MemorySearch Write Bash 5 20.0s 3.0%
fix performance issue Read Bash Edit 4 25.2s 3.0%
add error handling Read Bash Edit Attach 5 29.9s 3.0%
edit multiple functions Read Edit Bash 4 13.5s 2.9%
utility vague followup WebSearch WebFetch 3 16.6s 2.9%
social timezone page Write BrowserNavigate BrowserScreenshot Attach 4 17.3s 2.8%
no self limiting ToolSearch WebSearch WebFetch 4 16.3s 2.6%
conditional tool chain Read Edit Bash 4 11.3s 2.6%
fix index error Read Bash Edit 5 19.2s 2.5%
create test file Read Bash Write 5 21.3s 2.4%
utility weather now WebSearch WebFetch 3 17.0s 2.3%
read edit verify Read Edit Bash 4 11.2s 2.2%
code outline discovery ToolSearch CodeOutline 2 6.8s 2.0%
no describe just do Write Attach 2 7.0s 2.0%
utility release date WebSearch WebFetch 2 10.6s 1.9%
rung3 no claimant inline Read Bash 3 27.4s 1.9%
fix syntax error Bash Read Edit 4 9.1s 1.9%
social photo renamer Write Bash 2 15.6s 1.8%
bash chain with output Glob Write Bash 3 8.1s 1.8%
edit without read recovery Read Edit Bash 4 11.2s 1.7%
social pinnable rules Write BrowserNavigate BrowserScreenshot Attach 4 12.2s 1.6%
bash error recovery Bash Glob Write Attach 4 12.6s 1.4%
agent delegation discover ToolSearch 2 6.4s 1.3%
read and edit Read Edit Bash 4 8.7s 1.2%
memory guided refactor MemorySearch Read 2 9.5s 0.9%
utility screenshot a site BrowserNavigate BrowserScreenshot Attach 3 12.7s 0.9%
find then edit Grep Bash 2 7.3s 0.7%
brief responses Glob Bash 2 5.9s 0.7%
utility summarise a link WebFetch 1 3.8s 0.4%
social does it do what it says 0 10.7s 0.3%

Model openrouter/qwen/qwen3.7-flash. Every figure is a share of one Free day's budget — the same percentage /usage reports. Bars are scaled to that run's share against the heaviest in the set, so a heavier turn reads heavier. Runs scoring below 1.0 say so on the card rather than being dropped.

Home · Start here · Use it in a browser · Use it in Telegram · The system prompt behind these runs