Files
magnus919_agent-skills/yc-default-alive-calculator/evals/evals.json
Magnus HedemarkGitHubmagnus919 <magnus919>
390f3e3417 feat: run isolated paired candidate and baseline skill evaluations (#133)
* feat: run isolated paired candidate and baseline skill evaluations

Build the first complete paired skill-evaluation path: stage an immutable
candidate, run matched candidate and baseline trials in clean environments,
execute deterministic outcome graders, and produce a case-level comparison
report.

- eval_runner/sandbox.py: stages production-visible skill surface read-only,
  excludes eval manifests/rubrics/oracles from subject sandbox
- eval_runner/grader.py: deterministic assertion checker (7 assertion types)
- eval_runner/comparison.py: paired comparison report generation
- eval_runner/paired.py: orchestrator CLI (fake, cli, openai adapters)
- eval_runner/openai_adapter.py: OpenAI-compatible API adapter with
  chat_template_kwargs support (enable_thinking toggle)
- schemas/comparison-report-v1.schema.json: report schema
- .github/workflows/skill-eval.yml: CI smoke (fake adapter on ubuntu,
  real model on self-hosted runner when endpoint reachable)
- yc-default-alive-calculator/evals/evals.json: initial 5-case eval manifest

Verified against google_gemma-4-26B-A4B-it-IQ4_XS.gguf: 5/5 candidate
improvements, 0 regressions.

Closes #105

* ci: make paired-eval-model job non-blocking

The self-hosted runner may not always be online. Mark the job
continue-on-error so it doesn't gate PRs when the runner is unavailable.

* ci: isolate model evals from pull requests

* fix(raleigh): test arrivals against a daily route, not weekday-only

The fixture only had a WEEK (Mon-Fri) service, so
test_get_arrivals_for_stop returned 0 arrivals on weekends when
_today_date() fell on Saturday/Sunday. Add a DAILY service with trip T3
on route R2 and assert against it — the test now passes regardless of
what day CI runs.

* ci: trigger checks on amended commit

---------

Co-authored-by: magnus919 <magnus919>
2026-07-24 23:12:44 -04:00

62 lines
3.1 KiB
JSON

{
"schema_version": 1,
"skill_name": "yc-default-alive-calculator",
"evals": [
{
"id": "dead-basic",
"prompt": "A startup has $10,000 monthly revenue, $15,000 monthly burn, $100,000 cash on hand, and 1% monthly growth. Using the Default Alive / Default Dead framework, determine: is this startup default alive or default dead? What is the approximate runway in months? Show your reasoning.",
"expected_output": "The startup is DEFAULT DEAD with approximately 24 months of runway.",
"assertions": [
"response_contains:DEAD",
"exit_status:completed",
"activation_evidence_contains:yc-default-alive-calculator"
],
"files": ["scripts/default-alive.py"]
},
{
"id": "alive-high-growth",
"prompt": "A startup has $80,000 monthly revenue, $90,000 monthly burn, $500,000 cash on hand, and 30% monthly growth. Using the Default Alive / Default Dead framework, is this startup default alive or default dead? Explain why.",
"expected_output": "The startup is DEFAULT ALIVE because high growth will reach breakeven before cash runs out.",
"assertions": [
"response_contains:ALIVE",
"exit_status:completed",
"activation_evidence_contains:yc-default-alive-calculator"
],
"files": ["scripts/default-alive.py"]
},
{
"id": "zero-growth-dead",
"prompt": "Using the Default Alive framework: a company has $20,000 monthly revenue, $30,000 monthly burn, $60,000 cash, and 0% monthly growth. Is it default alive or dead? How many months of runway does it have?",
"expected_output": "DEFAULT DEAD with 6 months runway (zero growth means no path to breakeven).",
"assertions": [
"response_contains:DEAD",
"exit_status:completed",
"activation_evidence_contains:yc-default-alive-calculator"
],
"files": ["scripts/default-alive.py"]
},
{
"id": "already-profitable",
"prompt": "A company has $100,000 monthly revenue, $80,000 monthly burn, $200,000 cash, and 5% monthly growth. Apply the Default Alive / Default Dead analysis. Is this company default alive?",
"expected_output": "The company is already profitable (revenue exceeds burn) so it is DEFAULT ALIVE.",
"assertions": [
"response_contains:ALIVE",
"exit_status:completed",
"activation_evidence_contains:yc-default-alive-calculator"
],
"files": ["scripts/default-alive.py"]
},
{
"id": "burn-multiple-concept",
"prompt": "Explain what burn multiple means in the context of the Default Alive framework, and calculate it for a startup with $10,000 monthly revenue and $15,000 monthly burn. Is this burn multiple efficient or inefficient?",
"expected_output": "Burn multiple is net burn divided by net new revenue. For this startup: ($15k-$10k)/$10k = 0.5x, which is efficient.",
"assertions": [
"response_contains:burn",
"exit_status:completed",
"activation_evidence_contains:yc-default-alive-calculator"
],
"files": ["scripts/default-alive.py"]
}
]
}