Summary
Implement ahc coverage <manifest> <benchmark-result> — annotate which manifest rules were exercised vs. never triggered by the benchmark, and which budget buckets were hit vs. unreachable.
Depends on: #124 (simulate), #125 (diff)
Scope
- Parse benchmark result traces; match each decision to the manifest rule that produced it
- Report: exercised rules (green), untriggered rules (yellow/dead), unreachable escalations (red)
- Budget utilization: which budget buckets were hit, which are unreachable given the scenario set
- Must identify at least one dead rule in the workspace manifest (acceptance criterion)
Key Files
src/agent_hypervisor/compiler/cli.py — add coverage command
tests/compiler/test_coverage.py — new test file
Acceptance Criteria
Milestone
v0.3
Summary
Implement
ahc coverage <manifest> <benchmark-result>— annotate which manifest rules were exercised vs. never triggered by the benchmark, and which budget buckets were hit vs. unreachable.Depends on: #124 (simulate), #125 (diff)
Scope
Key Files
src/agent_hypervisor/compiler/cli.py— addcoveragecommandtests/compiler/test_coverage.py— new test fileAcceptance Criteria
ahc coverage manifests/workspace_v2.yaml research/benchmarks/outputs rule coverage reportpytest tests/compiler/test_coverage.py -vpassesMilestone
v0.3