Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
70 changes: 49 additions & 21 deletions .github/workflows/ci.yml
Original file line number Diff line number Diff line change
Expand Up @@ -6,42 +6,70 @@ permissions:
contents: read
jobs:
test:
name: Example tests
name: Example tests (${{ matrix.example.name }})
runs-on: ubuntu-latest
strategy:
matrix:
example:
- { name: prompt-technique-tournament, tests: eval/ }
- { name: latency-tuner, tests: sim/ }
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v5
with: { python-version: "3.12" }
- run: pip install -r examples/prompt-technique-tournament/requirements.txt
- run: pip install -r examples/${{ matrix.example.name }}/requirements.txt
- name: Run example + guardrail corpus
working-directory: examples/prompt-technique-tournament
run: python -m pytest eval/ -q --no-header
working-directory: examples/${{ matrix.example.name }}
run: python -m pytest ${{ matrix.example.tests }} -q --no-header
guardrail-integrity:
name: Guardrail integrity
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Hooks parse
run: |
bash -n template/.claude/hooks/guardrail.sh
bash -n examples/prompt-technique-tournament/.claude/hooks/guardrail.sh
- name: Floor 2 present and identical across template + example
for hook in template/.claude/hooks/guardrail.sh \
examples/prompt-technique-tournament/.claude/hooks/guardrail.sh \
examples/latency-tuner/.claude/hooks/guardrail.sh; do
bash -n "$hook"
done
- name: Floor 2 present and identical across template + examples
run: |
grep -qF 'Floor 2 — MACHINE SAFETY' template/.claude/hooks/guardrail.sh
grep -qF 'Floor 2 — MACHINE SAFETY' examples/prompt-technique-tournament/.claude/hooks/guardrail.sh
diff <(sed -n '/Floor 2/,$p' template/.claude/hooks/guardrail.sh) \
<(sed -n '/Floor 2/,$p' examples/prompt-technique-tournament/.claude/hooks/guardrail.sh)
for hook in template/.claude/hooks/guardrail.sh \
examples/prompt-technique-tournament/.claude/hooks/guardrail.sh \
examples/latency-tuner/.claude/hooks/guardrail.sh; do
grep -qF 'Floor 2 — MACHINE SAFETY' "$hook"
diff <(sed -n '/Floor 2/,$p' template/.claude/hooks/guardrail.sh) \
<(sed -n '/Floor 2/,$p' "$hook")
done
- name: Hooks wired for Bash AND write tools
run: |
grep -q 'Bash|Edit|Write|MultiEdit' template/.claude/settings.json
grep -q 'Bash|Edit|Write|MultiEdit' examples/prompt-technique-tournament/.claude/settings.json
- name: Wrappers parse
for settings in template/.claude/settings.json \
examples/prompt-technique-tournament/.claude/settings.json \
examples/latency-tuner/.claude/settings.json; do
grep -q 'Bash|Edit|Write|MultiEdit' "$settings"
done
- name: Wrappers and sandbox runner parse
run: |
bash -n template/run-iteration.sh
bash -n examples/prompt-technique-tournament/run-iteration.sh
bash -n template/drive-to-goal.sh
bash -n examples/prompt-technique-tournament/drive-to-goal.sh
- name: Runners identical across template + example
for script in template/run-iteration.sh template/drive-to-goal.sh \
template/sandbox/run-sandboxed.sh template/tools/metric-contract.sh \
examples/prompt-technique-tournament/run-iteration.sh \
examples/prompt-technique-tournament/drive-to-goal.sh \
examples/latency-tuner/run-iteration.sh \
examples/latency-tuner/drive-to-goal.sh; do
bash -n "$script"
done
- name: Runners identical across template + examples
run: |
diff template/run-iteration.sh examples/prompt-technique-tournament/run-iteration.sh
diff template/drive-to-goal.sh examples/prompt-technique-tournament/drive-to-goal.sh
for example in prompt-technique-tournament latency-tuner; do
diff template/run-iteration.sh "examples/$example/run-iteration.sh"
diff template/drive-to-goal.sh "examples/$example/drive-to-goal.sh"
done
- name: Shared tools identical across template + examples
run: |
for example in prompt-technique-tournament latency-tuner; do
diff template/tools/verdict.py "examples/$example/tools/verdict.py"
diff template/tools/metric-contract.sh "examples/$example/tools/metric-contract.sh"
done
- name: Mechanical gate self-test (the known-nothing control)
run: python3 template/tools/verdict.py self-test
Loading
Loading