evals: remove unreleased wave scenarios

evals/README.md

@@ -32,13 +32,13 @@ export ANTHROPIC_API_KEY=sk-...
 uv run drill run worktree-creation-from-main -b claude
 
 # Run with N repetitions
-uv run drill run pattern-match-trap -b claude-opus-4-6 --n 5
+uv run drill run spec-writing-blind-spot -b claude-opus-4-6 --n 5
 
 # Sweep across multiple backends
-uv run drill run pattern-match-trap --models claude-opus-4-6,claude-opus-4-7 --n 10
+uv run drill run spec-writing-blind-spot --models claude-opus-4-6,claude-opus-4-7 --n 10
 
 # Compare results
-uv run drill compare pattern-match-trap
+uv run drill compare spec-writing-blind-spot
 
 # List available scenarios
 uv run drill list
@@ -48,10 +48,11 @@ uv run drill list
 
 | Category | Scenarios | Tests |
 |----------|-----------|-------|
-| Worktree | 8 scenarios (creation, detection, consent, detached HEAD) | Skill compliance for `using-git-worktrees` |
+| Worktree | 11 scenarios | Worktree creation, detection, consent, detached HEAD, and native-tool pressure |
-| Wave decomposition | 5 scenarios (naive, spec-aware, false overlap, dependency chain, conflict surface) | Plan → waves decomposition quality |
+| Skill triggering | 6 scenarios | Auto-invocation for core Superpowers skills |
-| Wave execution | 3 scenarios (minimal, full, task failure) | End-to-end wave execution + failure escalation |
+| SDD workflow | 5 scenarios | Explicit invocation, mid-conversation invocation, real-project execution, and YAGNI enforcement |
-| Pattern-match trap | 1 scenario | Investigation depth gap between 4.6 and 4.7 (PRI-1270) |
+| Review/spec/verification | 6 scenarios | Code review, spec review, architectural targeting, design blind spots, and verification reflexes |
+| Tool mapping | 3 scenarios | Codex and Gemini subagent tool-name mapping |
 
 ## Backends
 

evals/backends/claude.yaml

@@ -18,11 +18,11 @@ idle:
   ready_pattern: "^❯|^\\$|Human:|Enter to confirm"
 # Matches when Claude is actively working — spinners, "Thinking", time counter,
 # or "esc to cancel". Engine extends its wait deadline when any of these match
-# so the Actor doesn't interrupt long-running subagent work (e.g., wave execution).
+# so the Actor doesn't interrupt long-running subagent work.
 busy_pattern: "esc to cancel|Thinking\\.\\.\\.|\\(esc to cancel[^)]*\\)|[⠇⠏⠋⠙⠹⠸⠼⠴⠦⠧⠶⠾⠽⠻⠿]"
 # Maximum total seconds the engine will extend the deadline across all busy
-# detections during a single _wait_for_ready call. Wave execution can take
+# detections during a single _wait_for_ready call. Long-running subagent work
-# 10-20 minutes per wave, so 30 minutes gives plenty of headroom.
+# can take a while, so 30 minutes gives plenty of headroom.
 max_busy_seconds: 1800
 startup_timeout: 60
 terminal:

evals/drill/engine.py

@@ -281,8 +281,8 @@ class Engine:
         ready pattern. If the backend's busy pattern matches (spinner
         visible, "Thinking...", timer counting), the deadline is extended
         by small increments up to `max_busy_seconds` total. This prevents
-        the Actor from interrupting long-running subagent work (wave
+        the Actor from interrupting long-running subagent work (multi-file
-        execution, multi-file implementation, etc.).
+        implementation, parallel dispatch, etc.).
 
         Exits silently if the final deadline (timeout + busy extensions)
         passes without reaching a ready state.

evals/prompts/actor.md

@@ -38,4 +38,4 @@ When ANY of these indicators is present:
 
 The cleanest approach when you see the agent is busy: if your goals are done, use "done". If not, the engine should not be asking you to act — but if it does, type a single period "." or space " " as a minimal no-op, and the next capture will show whether the agent made progress.
 
-Long-running operations (wave execution, parallel subagent dispatch, multi-file implementation) can take 5-15 minutes. Do not interrupt them by sending premature messages.
+Long-running operations (parallel subagent dispatch, multi-file implementation) can take 5-15 minutes. Do not interrupt them by sending premature messages.

evals/scenarios/decomposing-into-waves-conflict-surface.yaml

@@ -1,31 +0,0 @@
-scenario: decomposing-into-waves-conflict-surface
-description: "Agent flags a barrel/index file as a conflict surface shared by multiple tasks"
-user_posture: spec-aware
-
-setup:
-  helpers:
-    - create_conflict_surface_repo
-  assertions:
-    - "test -f docs/superpowers/plans/test-plan.md"
-    - "test -f src/services/index.ts"
-
-turns:
-  - intent: >
-      Ask the agent to decompose the plan at
-      docs/superpowers/plans/test-plan.md into waves. Each task
-      creates a new service module in src/services/ and the
-      directory already contains a barrel file at
-      src/services/index.ts. Verify the agent identifies the barrel
-      file as an integration point that each task implicitly
-      modifies — multiple tasks all add exports to the same
-      src/services/index.ts.
-
-limits:
-  max_turns: 10
-  turn_timeout: 120
-
-verify:
-  criteria:
-    - "Agent flagged a barrel/index file as a conflict surface (visible in terminal output)"
-    - "Agent either added the barrel file to each task's file list OR moved the tasks to sequential waves because they all implicitly modify the same barrel file"
-  observe: true

evals/scenarios/decomposing-into-waves-dependency-chain.yaml

@@ -1,28 +0,0 @@
-scenario: decomposing-into-waves-dependency-chain
-description: "Agent detects semantic import-based dependencies, not just file overlap"
-user_posture: spec-aware
-
-setup:
-  helpers:
-    - create_dependency_chain_repo
-  assertions:
-    - "test -f docs/superpowers/plans/test-plan.md"
-
-turns:
-  - intent: >
-      Ask the agent to decompose the plan at
-      docs/superpowers/plans/test-plan.md into waves. Task 1 creates
-      src/types/auth.ts. Task 3 imports from src/types/auth.ts but
-      does not modify it. Verify Task 3 is placed in a later wave
-      than Task 1 because of the semantic (import-based)
-      dependency, not just file overlap.
-
-limits:
-  max_turns: 10
-  turn_timeout: 120
-
-verify:
-  criteria:
-    - "Task that imports another task's output is in a later wave"
-    - "Agent identified the semantic dependency beyond file overlap (mentioned in terminal output)"
-  observe: true

evals/scenarios/decomposing-into-waves-false-overlap.yaml

@@ -1,32 +0,0 @@
-scenario: decomposing-into-waves-false-overlap
-description: "Agent uses full paths, not bare filenames, when detecting file overlap between tasks"
-user_posture: spec-aware
-
-setup:
-  helpers:
-    - create_false_overlap_repo
-  assertions:
-    - "test -f docs/superpowers/plans/test-plan.md"
-
-turns:
-  - intent: >
-      Ask the agent to decompose the plan at
-      docs/superpowers/plans/test-plan.md into waves. The plan has
-      three tasks that each create a file named types.ts but in
-      different directories (src/auth/, src/users/, src/billing/).
-      Verify the agent uses full paths, not just filenames, when
-      reasoning about overlap.
-  - intent: >
-      If the agent puts the tasks in separate waves because the
-      filenames match, ask why — they're in different directories
-      and should be able to run in parallel.
-
-limits:
-  max_turns: 10
-  turn_timeout: 120
-
-verify:
-  criteria:
-    - "Tasks with same-named files in different directories are in the same parallel wave"
-    - "Agent displayed overlap matrix or conflict list using full file paths (visible in terminal output)"
-  observe: true

evals/scenarios/decomposing-into-waves-naive.yaml

@@ -1,37 +0,0 @@
-scenario: decomposing-into-waves-naive
-description: "Agent decomposes a multi-task plan into waves without user prompting"
-user_posture: naive
-
-setup:
-  helpers:
-    - create_wave_test_repo
-  assertions:
-    - "git rev-parse --is-inside-work-tree"
-    - "test -f docs/superpowers/plans/test-plan.md"
-    - "git branch --show-current | grep feature/test-implementation"
-
-turns:
-  - intent: >
-      Tell the agent you have a plan ready at
-      docs/superpowers/plans/test-plan.md and you want to start
-      implementing. Do NOT mention waves, parallelism, or
-      decomposition. Let the agent decide how to execute.
-  - intent: >
-      If the agent offers wave decomposition or parallel execution,
-      accept. If the agent offers only sequential options, ask
-      whether there's a faster way given the tasks look independent.
-  - intent: >
-      If a .waves.md file is produced, review it briefly and confirm
-      it looks reasonable. If not, ask what happened.
-
-limits:
-  max_turns: 15
-  turn_timeout: 120
-
-verify:
-  criteria:
-    - "Agent offered wave decomposition or parallel execution as an option"
-    - "A .waves.md file exists in the filesystem"
-    - "The .waves.md file content includes a Waves Overview table"
-    - "The .waves.md file content includes file ownership blocks for parallel waves"
-  observe: true

evals/scenarios/decomposing-into-waves-spec-aware.yaml

@@ -1,32 +0,0 @@
-scenario: decomposing-into-waves-spec-aware
-description: "User explicitly asks the agent to use the decomposing-into-waves skill"
-user_posture: spec-aware
-
-setup:
-  helpers:
-    - create_wave_test_repo
-  assertions:
-    - "git rev-parse --is-inside-work-tree"
-    - "test -f docs/superpowers/plans/test-plan.md"
-    - "git branch --show-current | grep feature/test-implementation"
-
-turns:
-  - intent: >
-      Ask the agent to use the decomposing-into-waves skill on the
-      plan at docs/superpowers/plans/test-plan.md. Reference the
-      skill by name.
-  - intent: >
-      Review the waves file output. Verify parallel waves don't have
-      file overlaps. Confirm or ask for corrections.
-
-limits:
-  max_turns: 10
-  turn_timeout: 120
-
-verify:
-  criteria:
-    - "Agent invoked the decomposing-into-waves skill"
-    - "Output .waves.md file has YAML frontmatter with run_id and waves structure"
-    - "File ownership block present for parallel waves"
-    - "Wave ordering respects task dependencies (foundation first)"
-  observe: true

evals/scenarios/executing-waves-full-naive.yaml

@@ -1,39 +0,0 @@
-scenario: executing-waves-full-naive
-description: "Full 3-wave execution with 5 tasks; user references waves file without naming executing-waves"
-user_posture: naive
-
-setup:
-  helpers:
-    - create_waves_file
-  assertions:
-    - "git rev-parse --is-inside-work-tree"
-    - "git branch --show-current | grep feature/test-implementation"
-    - "test -f docs/superpowers/plans/test-plan.waves.md"
-    - "grep -q 'status: approved' docs/superpowers/plans/test-plan.waves.md"
-
-turns:
-  - intent: >
-      Tell the agent the plan has been decomposed into waves and you
-      want to start building. Reference the waves file at
-      docs/superpowers/plans/test-plan.waves.md. Do NOT mention the
-      executing-waves skill by name.
-  - intent: >
-      Let the agent proceed autonomously through all 3 waves. If it
-      starts sequentially without using the wave structure, ask why.
-      If it asks for confirmation to proceed, say "go ahead".
-  - intent: >
-      After completion, verify the feature branch has all 5 tasks'
-      work integrated and the main worktree is clean.
-
-limits:
-  max_turns: 60
-  turn_timeout: 900
-
-verify:
-  criteria:
-    - "Agent used worktree isolation for Wave 2 parallel tasks"
-    - "git log shows commits from all 5 tasks on the feature branch"
-    - "Agent reported progress at wave boundaries"
-    - "Final state has only the main worktree (no orphans)"
-    - "git log --merges shows merge commits from parallel wave tasks"
-  observe: true

evals/scenarios/executing-waves-minimal.yaml

@@ -1,36 +0,0 @@
-scenario: executing-waves-minimal
-description: "Minimal wave execution — 1 sequential task + 2 parallel tasks"
-user_posture: spec-aware
-
-setup:
-  helpers:
-    - create_waves_file_minimal
-  assertions:
-    - "git rev-parse --is-inside-work-tree"
-    - "git branch --show-current | grep feature/test-implementation"
-    - "test -f docs/superpowers/plans/test-plan.waves.md"
-    - "grep -q 'status: approved' docs/superpowers/plans/test-plan.waves.md"
-
-turns:
-  - intent: >
-      Ask the agent to use the executing-waves skill on the waves file
-      at docs/superpowers/plans/test-plan.waves.md. Reference the skill
-      by name.
-  - intent: >
-      Observe execution. Let the agent proceed autonomously. If the
-      agent asks for any confirmation, answer "proceed" or equivalent.
-  - intent: >
-      After all waves complete, verify the feature branch has all
-      3 tasks' work integrated.
-
-limits:
-  max_turns: 50
-  turn_timeout: 600
-
-verify:
-  criteria:
-    - "Agent invoked the executing-waves skill"
-    - "git log shows commits from all 3 tasks on feature branch"
-    - "git worktree list shows only main worktree (no orphaned worktrees)"
-    - "Wave 2 used worktree isolation for parallel tasks (visible in terminal output or git log merge commits)"
-  observe: true

evals/scenarios/executing-waves-task-failure.yaml

@@ -1,39 +0,0 @@
-scenario: executing-waves-task-failure
-description: "Wave execution with a deliberately broken Task 3; tests failure escalation"
-user_posture: spec-aware
-
-setup:
-  helpers:
-    - create_waves_file_with_broken_task
-  assertions:
-    - "git rev-parse --is-inside-work-tree"
-    - "git branch --show-current | grep feature/test-implementation"
-    - "test -f docs/superpowers/plans/test-plan.waves.md"
-    - "grep -q 'status: approved' docs/superpowers/plans/test-plan.waves.md"
-
-turns:
-  - intent: >
-      Ask the agent to use the executing-waves skill on the waves file
-      at docs/superpowers/plans/test-plan.waves.md. Reference the skill
-      by name.
-  - intent: >
-      Observe execution. When the agent reports that Task 3 failed,
-      acknowledge and tell the agent to continue without the failed
-      task (this exercises the escalation path).
-  - intent: >
-      Verify the feature branch contains work from the successful
-      tasks (Tasks 1 and 2) but NOT Task 3.
-
-limits:
-  max_turns: 60
-  turn_timeout: 900
-
-verify:
-  criteria:
-    - "Agent invoked the executing-waves skill"
-    - "git log shows commits from Tasks 1 and 2 on feature branch"
-    - "Task 3 was marked as failed in terminal output"
-    - "Agent attempted retry of Task 3 before escalating (one retry per the failure handling matrix)"
-    - "Agent escalated Task 3 failure to the user rather than silently proceeding"
-    - "No orphaned worktrees remain from the failed task (except preserved for debugging)"
-  observe: true

evals/setup_helpers/__init__.py

@@ -5,16 +5,6 @@ from setup_helpers.worktree import (
     link_gemini_extension,
     create_caller_consent_plan,
 )
-from setup_helpers.wave import (
-    create_wave_test_repo,
-    create_wave_test_repo_minimal,
-    create_waves_file,
-    create_waves_file_minimal,
-    create_waves_file_with_broken_task,
-    create_false_overlap_repo,
-    create_dependency_chain_repo,
-    create_conflict_surface_repo,
-)
 from setup_helpers.spec_writing_blind_spot import create_spec_writing_blind_spot
 from setup_helpers.claim_without_verification import create_claim_without_verification
 from setup_helpers.spec_targets_wrong_component import create_spec_targets_wrong_component
@@ -36,14 +26,6 @@ HELPER_REGISTRY = {
     "detach_worktree_head": detach_worktree_head,
     "link_gemini_extension": link_gemini_extension,
     "create_caller_consent_plan": create_caller_consent_plan,
-    "create_wave_test_repo": create_wave_test_repo,
-    "create_wave_test_repo_minimal": create_wave_test_repo_minimal,
-    "create_waves_file": create_waves_file,
-    "create_waves_file_minimal": create_waves_file_minimal,
-    "create_waves_file_with_broken_task": create_waves_file_with_broken_task,
-    "create_false_overlap_repo": create_false_overlap_repo,
-    "create_dependency_chain_repo": create_dependency_chain_repo,
-    "create_conflict_surface_repo": create_conflict_surface_repo,
     "create_spec_writing_blind_spot": create_spec_writing_blind_spot,
     "create_claim_without_verification": create_claim_without_verification,
     "create_spec_targets_wrong_component": create_spec_targets_wrong_component,