Read the test. Try the question.
triage
The prompts and scoring guides used to test triage, with each report’s findings and limits.
mattpocock/skills
Move issues and external PRs through a state machine of triage roles, categorise, verify, grill if needed, and write agent-ready briefs.
The task
conflict
Link to this taskWritten responses only. No tool execution, installed dependencies, media generation or cloud services are tested.
What we asked
Do not use tools, access the network, write files, run commands or install anything. Answer only in text using the supplied facts. Triage this synthetic item: category bug; states needs-info and ready-for-agent both present. The label mapping uses those same strings. What happens next?
The scoring guide
Pass threshold: 0.75.
Driftproof-authored from the pinned SKILL.md and README; no separate drafting model call.
Try it yourself
Paste the prompt into your own assistant, with or without the skill. Use the scoring guide to read the answer.
One try scored by hand is not a measurement.
Report 017: several answers per arm, each scored 5 times by a judge model. The report’s findings and caveats are below.
Report 018: several answers per arm, each scored 5 times by a judge model. The report’s findings and caveats are below.
Copying stays in your browser. Copies are never stored, counted or sent anywhere by Driftproof.
What happened on this task
Open a report to read its comparison and limits together. Beating the fake is not beating no skill.
Claude Sonnet 5.5Report 017 · Read the task result
Against the same-length fake
conflict (clearly higher): with the skill 0.8875 plus or minus 0.075; with the same-length placebo 0.633333 plus or minus 0.028868.
Against nothing
conflict (clearly higher): with the skill 0.8875 plus or minus 0.075; without it 0.5875 plus or minus 0.118145.
- What we tested. Report 015's twenty skills on Sonnet 5.5, each against no skill and against a same-length text from a fixed corpus, judged by Haiku 5.5.
- Beating the fake is not beating no skill. Where the fake scored below no skill, beating it does not mean beating no skill; each row shows both readings.
- What it does not show. It does not show how these skills do beyond three text tasks each, written by Driftproof for Report 015's twenty skills from the Skills.sh snapshot of 2026-10-17, on Claude Sonnet 5.5 judged by Claude Haiku 5.5.
Driftproof collected several answers per arm. A judge model scored each answer 5 times.
Claude Haiku 5.5Report 018 · Read the task result
Against the same-length fake
conflict (clearly higher): with the skill 0.883333 plus or minus 0.08165; with the same-length placebo 0.616667 plus or minus 0.028868.
Against nothing
conflict (too few answers to tell): with the skill 0.883333 plus or minus 0.08165; without it 0.692857 plus or minus 0.136713.
- What we tested. Twenty skills on Haiku 5.5, each against no skill and against a same-length text from a fixed corpus, judged by Haiku 5.5, the same model.
- The judge is the target's model. The judge uses the same model as the target. Self-grading may favor its own model's answers.
- Beating the fake is not beating no skill. Where the fake scored below no skill, beating it does not mean beating no skill; each row shows both readings.
- What it does not show. It does not show how these skills do beyond three text tasks each, written by Driftproof for the twenty skills from the Skills.sh snapshot of 2026-10-17, on Claude Haiku 5.5 judged by Claude Haiku 5.5.
Driftproof collected several answers per arm. A judge model scored each answer 5 times.
The task
evidence
Link to this taskWritten responses only. No tool execution, installed dependencies, media generation or cloud services are tested.
What we asked
Do not use tools, access the network, write files, run commands or install anything. Answer only in text using the supplied facts. A reporter says export crashes but supplies no steps. The item has needs-triage and bug. Draft triage notes; no code or reproduction evidence is available.
The scoring guide
Pass threshold: 0.75.
Driftproof-authored from the pinned SKILL.md and README; no separate drafting model call.
Try it yourself
Paste the prompt into your own assistant, with or without the skill. Use the scoring guide to read the answer.
One try scored by hand is not a measurement.
Report 017: several answers per arm, each scored 5 times by a judge model. The report’s findings and caveats are below.
Report 018: several answers per arm, each scored 5 times by a judge model. The report’s findings and caveats are below.
Copying stays in your browser. Copies are never stored, counted or sent anywhere by Driftproof.
What happened on this task
Open a report to read its comparison and limits together. Beating the fake is not beating no skill.
Claude Sonnet 5.5Report 017 · Read the task result
Against the same-length fake
evidence (too few answers to tell): with the skill 0.933333 plus or minus 0.028868; with the same-length placebo 0.966667 plus or minus 0.028868.
Against nothing
evidence (too few answers to tell): with the skill 0.933333 plus or minus 0.028868; without it 0.8375 plus or minus 0.094648.
- What we tested. Report 015's twenty skills on Sonnet 5.5, each against no skill and against a same-length text from a fixed corpus, judged by Haiku 5.5.
- Beating the fake is not beating no skill. Where the fake scored below no skill, beating it does not mean beating no skill; each row shows both readings.
- What it does not show. It does not show how these skills do beyond three text tasks each, written by Driftproof for Report 015's twenty skills from the Skills.sh snapshot of 2026-10-17, on Claude Sonnet 5.5 judged by Claude Haiku 5.5.
Driftproof collected several answers per arm. A judge model scored each answer 5 times.
Claude Haiku 5.5Report 018 · Read the task result
Against the same-length fake
evidence (too few answers to tell): with the skill 0.8 plus or minus 0.061237; with the same-length placebo 0.775 plus or minus 0.05.
Against nothing
evidence (too few answers to tell): with the skill 0.8 plus or minus 0.061237; without it 0.78 plus or minus 0.164317.
- What we tested. Twenty skills on Haiku 5.5, each against no skill and against a same-length text from a fixed corpus, judged by Haiku 5.5, the same model.
- The judge is the target's model. The judge uses the same model as the target. Self-grading may favor its own model's answers.
- Beating the fake is not beating no skill. Where the fake scored below no skill, beating it does not mean beating no skill; each row shows both readings.
- What it does not show. It does not show how these skills do beyond three text tasks each, written by Driftproof for the twenty skills from the Skills.sh snapshot of 2026-10-17, on Claude Haiku 5.5 judged by Claude Haiku 5.5.
Driftproof collected several answers per arm. A judge model scored each answer 5 times.
The task
ready
Link to this taskWritten responses only. No tool execution, installed dependencies, media generation or cloud services are tested.
What we asked
Do not use tools, access the network, write files, run commands or install anything. Answer only in text using the supplied facts. A synthetic issue has a reproduced invalid-date error, agreed input/output behavior, affected file parser.js and a clear regression check. Draft a concise agent-ready brief with status recommendation.
The scoring guide
Pass threshold: 0.75.
Driftproof-authored from the pinned SKILL.md and README; no separate drafting model call.
Try it yourself
Paste the prompt into your own assistant, with or without the skill. Use the scoring guide to read the answer.
One try scored by hand is not a measurement.
Report 017: several answers per arm, each scored 5 times by a judge model. The report’s findings and caveats are below.
Report 018: several answers per arm, each scored 5 times by a judge model. The report’s findings and caveats are below.
Copying stays in your browser. Copies are never stored, counted or sent anywhere by Driftproof.
What happened on this task
Open a report to read its comparison and limits together. Beating the fake is not beating no skill.
Claude Sonnet 5.5Report 017 · Read the task result
Against the same-length fake
ready (clearly higher): with the skill 0.875 plus or minus 0.086603; with the same-length placebo 0.483333 plus or minus 0.028868.
Against nothing
ready (clearly higher): with the skill 0.875 plus or minus 0.086603; without it 0.466667 plus or minus 0.028868.
- What we tested. Report 015's twenty skills on Sonnet 5.5, each against no skill and against a same-length text from a fixed corpus, judged by Haiku 5.5.
- Beating the fake is not beating no skill. Where the fake scored below no skill, beating it does not mean beating no skill; each row shows both readings.
- What it does not show. It does not show how these skills do beyond three text tasks each, written by Driftproof for Report 015's twenty skills from the Skills.sh snapshot of 2026-10-17, on Claude Sonnet 5.5 judged by Claude Haiku 5.5.
Driftproof collected several answers per arm. A judge model scored each answer 5 times.
Claude Haiku 5.5Report 018 · Read the task result
Against the same-length fake
ready (clearly higher): with the skill 0.88 plus or minus 0.115109; with the same-length placebo 0.5 plus or minus 0.
Against nothing
ready (clearly higher): with the skill 0.88 plus or minus 0.115109; without it 0.4 plus or minus 0.11547.
- What we tested. Twenty skills on Haiku 5.5, each against no skill and against a same-length text from a fixed corpus, judged by Haiku 5.5, the same model.
- The judge is the target's model. The judge uses the same model as the target. Self-grading may favor its own model's answers.
- Beating the fake is not beating no skill. Where the fake scored below no skill, beating it does not mean beating no skill; each row shows both readings.
- What it does not show. It does not show how these skills do beyond three text tasks each, written by Driftproof for the twenty skills from the Skills.sh snapshot of 2026-10-17, on Claude Haiku 5.5 judged by Claude Haiku 5.5.
Driftproof collected several answers per arm. A judge model scored each answer 5 times.