Plan: Workspace obsolete test cleanup and de-risking

[graysurf]

Plan: Workspace obsolete test cleanup and de-risking

Overview

This plan defines a full-workspace cleanup pass for obsolete tests and stale test helpers across all crates under crates/*.
The goal is to remove tests that no longer protect behavior (for example dead-code-shield tests, deprecated-path leftovers, and unreferenced helper functions), while preserving contract fidelity for CLI output, warnings, color behavior, JSON shape, and exit semantics.
Execution is organized as sprint-level integration gates: inventory first, safe helper cleanup next, high-risk behavior cleanup after that, then prevention/guardrail hardening.
No cross-sprint execution parallelism is planned; parallelization is only used within each sprint.

Scope

• In scope:
  • All Rust test surfaces in this workspace (crates/*/tests, #[cfg(test)] modules, test fixtures tightly coupled to removed behavior).
  • Test helper modules such as tests/common.rs, tests/support/mod.rs, tests/real_common.rs, and similar scaffolding.
  • Obsolete #[allow(dead_code)] usage in tests when the underlying helper is now unused.
  • Test cleanup reporting artifacts under $AGENT_HOME/out/.
• Out of scope:
  • Product feature changes unrelated to test cleanup.
  • Intentional parity/contract tests that still guard user-visible behavior.
  • Major architecture migrations outside test surfaces.

Assumptions

1. Contract fidelity remains the top priority: no unapproved drift in human output text, warning style, color behavior, JSON contracts, or exit codes.
2. A test is considered obsolete only when it no longer guards reachable behavior or contractual compatibility.
3. Each sprint is a hard integration gate and must pass its validation commands before moving to the next sprint.
4. Temporary reports and inventories are written to $AGENT_HOME/out/workspace-test-cleanup/ (not as canonical repo docs).
5. Final delivery must pass required repo checks and workspace coverage gate from DEVELOPMENT.md.

Success criteria

• Every crate has an explicit stale-test classification record with rationale (remove, keep, rewrite, defer).
• Unused test helper functions/modules are removed or consolidated, and related dead-code allowances are minimized.
• Deprecated/removed-path tests are cleaned without reducing contract coverage.
• Required checks and workspace coverage remain green after cleanup.
• A repeatable stale-test audit workflow exists to prevent re-accumulation.

Task Decomposition

Task	Summary	Owner	Branch	Worktree	Execution Mode	PR	Status	Notes
S1T1	Build workspace stale-test scanner and normalized output schema	subagent-s1-t1	issue/s1-t1-build-workspace-stale-test-scanner-and-normalize	issue-s1-t1	per-sprint	#262	done	sprint=S1; plan-task:Task 1.1; validate=mkdir -p "$AGENT_HOME/out/workspace-test-cleanup"; pr-grouping=per-sprint; pr-group=s1; shared-pr-anchor=S1T1
S1T2	Build unused-helper call graph for test scaffolding modules	subagent-s1-t1	issue/s1-t1-build-workspace-stale-test-scanner-and-normalize	issue-s1-t1	per-sprint	#262	done	sprint=S1; plan-task:Task 1.2; deps=Task 1.1; validate=test -s "$AGENT_HOME/out/workspace-test-cleanup/helper-callgraph.tsv"; pr-grouping=per-sprint; pr-group=s1; shared-pr-anchor=S1T1
S1T3	Define stale-test decision rubric and contract-protection allowlist	subagent-s1-t1	issue/s1-t1-build-workspace-stale-test-scanner-and-normalize	issue-s1-t1	per-sprint	#262	done	sprint=S1; plan-task:Task 1.3; deps=Task 1.1; validate=test -s "$AGENT_HOME/out/workspace-test-cleanup/decision-rubric.md"; pr-grouping=per-sprint; pr-group=s1; shared-pr-anchor=S1T1
S1T4	Produce sprint execution manifest and crate-tier lanes	subagent-s1-t1	issue/s1-t1-build-workspace-stale-test-scanner-and-normalize	issue-s1-t1	per-sprint	#262	done	sprint=S1; plan-task:Task 1.4; deps=Task 1.2,Task 1.3; validate=test -s "$AGENT_HOME/out/workspace-test-cleanup/execution-manifest.md"; pr-grouping=per-sprint; pr-group=s1; shared-pr-anchor=S1T1
S2T1	Remove orphaned helper functions in low-risk crates	subagent-s2-t1	issue/s2-t1-remove-orphaned-helper-functions-in-low-risk-cra	issue-s2-t1	per-sprint	#263	done	sprint=S2; plan-task:Task 2.1; deps=Task 1.4; validate=cargo test -p nils-git-scope -p nils-image-processing -p nils-api-testing-core -p nils-git-summary -p nils-plan-tooling; pr-grouping=per-sprint; pr-group=s2; shared-pr-anchor=S2T1
S2T2	Reduce obsolete #[allow(dead_code)] and crate-level dead-code shields in test helpers	subagent-s2-t1	issue/s2-t1-remove-orphaned-helper-functions-in-low-risk-cra	issue-s2-t1	per-sprint	#263	done	sprint=S2; plan-task:Task 2.2; deps=Task 1.4; validate=cargo test -p nils-fzf-cli -p nils-semantic-commit -p nils-git-lock -p nils-memo-cli -p nils-agent-docs; pr-grouping=per-sprint; pr-group=s2; shared-pr-anchor=S2T1
S2T3	Consolidate nils-test-support helper primitives (atomic lane A)	subagent-s2-t1	issue/s2-t1-remove-orphaned-helper-functions-in-low-risk-cra	issue-s2-t1	per-sprint	#263	done	sprint=S2; plan-task:Task 2.3; deps=Task 2.1,Task 2.2; validate=cargo test -p nils-test-support; pr-grouping=per-sprint; pr-group=s2; shared-pr-anchor=S2T1
S2T4	Adopt consolidated helpers in medium-risk crates (atomic lane B)	subagent-s2-t1	issue/s2-t1-remove-orphaned-helper-functions-in-low-risk-cra	issue-s2-t1	per-sprint	#263	done	sprint=S2; plan-task:Task 2.4; deps=Task 2.3; validate=cargo test -p nils-git-cli -p nils-git-lock -p nils-semantic-commit -p nils-plan-issue-cli -p nils-macos-agent; pr-grouping=per-sprint; pr-group=s2; shared-pr-anchor=S2T1
S2T5	Tier-2 regression sweep and fallout fixes	subagent-s2-t1	issue/s2-t1-remove-orphaned-helper-functions-in-low-risk-cra	issue-s2-t1	per-sprint	#263	done	sprint=S2; plan-task:Task 2.5; deps=Task 2.4; validate=cargo test -p nils-git-cli -p nils-plan-issue-cli -p nils-macos-agent -p nils-screen-record -p nils-git-summary -p nils-fzf-cli; pr-grouping=per-sprint; pr-group=s2; shared-pr-anchor=S2T1
S3T1	Identify deprecated/removed-path tests and fixture dependencies	subagent-s3-t1	issue/s3-t1-identify-deprecated-removed-path-tests-and-fixtu	issue-s3-t1	per-sprint	#264	done	sprint=S3; plan-task:Task 3.1; deps=Task 2.5; validate=test -s "$AGENT_HOME/out/workspace-test-cleanup/deprecated-path-candidates.tsv"; pr-grouping=per-sprint; pr-group=s3; shared-pr-anchor=S3T1
S3T2	Remove obsolete deprecated-path tests and stale fixtures	subagent-s3-t1	issue/s3-t1-identify-deprecated-removed-path-tests-and-fixtu	issue-s3-t1	per-sprint	#264	done	sprint=S3; plan-task:Task 3.2; deps=Task 3.1; validate=cargo test --workspace; pr-grouping=per-sprint; pr-group=s3; shared-pr-anchor=S3T1
S3T3	Rebaseline parity and contract suites for current command surfaces	subagent-s3-t1	issue/s3-t1-identify-deprecated-removed-path-tests-and-fixtu	issue-s3-t1	per-sprint	#264	done	sprint=S3; plan-task:Task 3.3; deps=Task 3.2; validate=cargo test -p nils-codex-cli -p nils-gemini-cli -p nils-plan-issue-cli -p nils-api-rest -p nils-api-gql; pr-grouping=per-sprint; pr-group=s3; shared-pr-anchor=S3T1
S3T4	Full workspace regression gate after high-risk cleanup	subagent-s3-t1	issue/s3-t1-identify-deprecated-removed-path-tests-and-fixtu	issue-s3-t1	per-sprint	#264	done	sprint=S3; plan-task:Task 3.4; deps=Task 3.3; validate=./.agents/skills/nils-cli-verify-required-checks/scripts/nils-cli-verify-required-checks.sh; pr-grouping=per-sprint; pr-group=s3; shared-pr-anchor=S3T1
S4T1	Add strict CI stale-test audit command	subagent-s4-t1	issue/s4-t1-add-strict-ci-stale-test-audit-command	issue-s4-t1	per-sprint	#265	done	sprint=S4; plan-task:Task 4.1; deps=Task 3.4; validate=mkdir -p "$AGENT_HOME/out/workspace-test-cleanup"; pr-grouping=per-sprint; pr-group=s4; shared-pr-anchor=S4T1
S4T2	Document stale-test lifecycle and review checklist	subagent-s4-t1	issue/s4-t1-add-strict-ci-stale-test-audit-command	issue-s4-t1	per-sprint	#265	done	sprint=S4; plan-task:Task 4.2; deps=Task 3.4; validate=rg -n 'stale test/remove/keep/rewrite/defer/test-stale-audit' docs/runbooks/test-cleanup-governance.md DEVELOPMENT.md AGENTS.md; pr-grouping=per-sprint; pr-group=s4; shared-pr-anchor=S4T1
S4T3	Wire stale-test audit into required check flow and run final workspace gate	subagent-s4-t1	issue/s4-t1-add-strict-ci-stale-test-audit-command	issue-s4-t1	per-sprint	#265	done	sprint=S4; plan-task:Task 4.3; deps=Task 4.1,Task 4.2; validate=./.agents/skills/nils-cli-verify-required-checks/scripts/nils-cli-verify-required-checks.sh; pr-grouping=per-sprint; pr-group=s4; shared-pr-anchor=S4T1

Consistency Rules

• Status must be one of: planned, in-progress, blocked, done.
• Status = in-progress or done requires non-TBD execution metadata (Owner, Branch, Worktree, Execution Mode, PR).
• Owner must be a subagent identifier (contains subagent) once the task is assigned; main-agent ownership is invalid for implementation tasks.
• Execution Mode should be one of: per-sprint, pr-isolated, pr-shared (or TBD before assignment).
• Branch and Worktree uniqueness is enforced only for rows using Execution Mode = pr-isolated.

Risks / Uncertainties

• Sprint approvals may be recorded before final close; issue stays open until final plan acceptance.
• Close gate fails if task statuses or PR merge states in the issue body are incomplete.

Evidence

• Plan source: docs/plans/workspace-obsolete-tests-cleanup-plan.md
• Sprint approvals: issue comments (one comment per accepted sprint)
• Final approval: issue/pull comment URL passed to close-plan

[graysurf]

Sprint 1 Start

• Sprint: 1 (Inventory and stale-test decision graph)
• Tasks in sprint: 4
• Note: Main-agent starts this sprint on the plan issue and dispatches implementation to subagents.
• Execution Mode comes from current Task Decomposition for each sprint task.

Task	Summary	Execution Mode
S1T1	Build workspace stale-test scanner and normalized output schema	per-sprint
S1T2	Build unused-helper call graph for test scaffolding modules	per-sprint
S1T3	Define stale-test decision rubric and contract-protection allowlist	per-sprint
S1T4	Produce sprint execution manifest and crate-tier lanes	per-sprint

Sprint 1: Inventory and stale-test decision graph

Goal: Build a deterministic full-workspace stale-test inventory and executable dependency graph before deleting tests.
PR grouping intent: per-sprint.
Execution Profile: parallel-x2 (parallel width 2).
Demo/Validation:

• Command(s):
  • mkdir -p "$AGENT_HOME/out/workspace-test-cleanup"
  • bash scripts/dev/workspace-test-stale-audit.sh --format tsv --out "$AGENT_HOME/out/workspace-test-cleanup/stale-tests.tsv"
  • plan-tooling validate --file docs/plans/workspace-obsolete-tests-cleanup-plan.md
• Verify:
  • Inventory includes all crates and stale-candidate test entries with unique IDs.
  • Every candidate has a deterministic decision (remove|keep|rewrite|defer) and rationale.
    Sprint scorecard:
• TotalComplexity: 16
• CriticalPathComplexity: 13
• MaxBatchWidth: 2
• OverlapHotspots: scripts/dev/workspace-test-stale-audit.sh, $AGENT_HOME/out/workspace-test-cleanup/stale-tests.tsv, $AGENT_HOME/out/workspace-test-cleanup/decision-matrix.md.
  Parallelizable tasks:
• Task 1.2 and Task 1.3 can run in parallel after Task 1.1.

Task 1.1: Build workspace stale-test scanner and normalized output schema

• Location:
  • scripts/dev/workspace-test-stale-audit.sh
  • $AGENT_HOME/out/workspace-test-cleanup/stale-tests.tsv
  • $AGENT_HOME/out/workspace-test-cleanup/stale-tests.md
• Description: Create a deterministic scanner that indexes test modules, helper functions, #[allow(dead_code)] usage, helper-call fanout, and deprecated-path markers per crate.
• Dependencies:
  • none
• Complexity: 5
• Complexity notes: New audit script plus schema stability requirements across reruns.
• Acceptance criteria:
  • scripts/dev/workspace-test-stale-audit.sh is created in this task if absent.
  • Output includes every crate under crates/* exactly once in summary section.
  • Candidate rows include id, crate, path, symbol_or_test, signal, proposed_action, confidence.
  • Output is deterministic across two immediate reruns.
• Validation:
  • mkdir -p "$AGENT_HOME/out/workspace-test-cleanup"
  • bash scripts/dev/workspace-test-stale-audit.sh --format tsv --out "$AGENT_HOME/out/workspace-test-cleanup/stale-tests.tsv"
  • cp "$AGENT_HOME/out/workspace-test-cleanup/stale-tests.tsv" "$AGENT_HOME/out/workspace-test-cleanup/stale-tests.r1.tsv"
  • bash scripts/dev/workspace-test-stale-audit.sh --format tsv --out "$AGENT_HOME/out/workspace-test-cleanup/stale-tests.r2.tsv"
  • diff -u "$AGENT_HOME/out/workspace-test-cleanup/stale-tests.r1.tsv" "$AGENT_HOME/out/workspace-test-cleanup/stale-tests.r2.tsv"

Task 1.2: Build unused-helper call graph for test scaffolding modules

• Location:
  • $AGENT_HOME/out/workspace-test-cleanup/helper-callgraph.tsv
  • $AGENT_HOME/out/workspace-test-cleanup/helper-orphans.tsv
  • crates/git-scope/tests/common.rs
  • crates/fzf-cli/tests/common.rs
  • crates/image-processing/tests/common.rs
  • crates/api-testing-core/tests/support/mod.rs
  • crates/memo-cli/tests/support/mod.rs
  • crates/plan-tooling/tests/common.rs
• Description: Compute helper definition-to-callsite graph for test helper modules and identify provably unreferenced helpers suitable for safe removal.
• Dependencies:
  • Task 1.1
• Complexity: 4
• Complexity notes: Graph generation is straightforward, but false positives must be constrained.
• Acceptance criteria:
  • Every helper function in scanned helper modules has callsite count metadata.
  • Orphan list contains only helpers with no external callsites in same crate test tree.
  • Potential reflection/macro edge cases are flagged as manual-review.
• Validation:
  • test -s "$AGENT_HOME/out/workspace-test-cleanup/helper-callgraph.tsv"
  • test -s "$AGENT_HOME/out/workspace-test-cleanup/helper-orphans.tsv"
  • awk -F '\t' 'NR>1 && $4=="0" {print $1}' "$AGENT_HOME/out/workspace-test-cleanup/helper-callgraph.tsv" | sort > "$AGENT_HOME/out/workspace-test-cleanup/orphans-from-callgraph.ids"
  • awk -F '\t' 'NR>1 {print $1}' "$AGENT_HOME/out/workspace-test-cleanup/helper-orphans.tsv" | sort > "$AGENT_HOME/out/workspace-test-cleanup/orphans-reported.ids"
  • diff -u "$AGENT_HOME/out/workspace-test-cleanup/orphans-from-callgraph.ids" "$AGENT_HOME/out/workspace-test-cleanup/orphans-reported.ids"

Task 1.3: Define stale-test decision rubric and contract-protection allowlist

• Location:
  • $AGENT_HOME/out/workspace-test-cleanup/decision-rubric.md
  • $AGENT_HOME/out/workspace-test-cleanup/contract-allowlist.tsv
  • AGENTS.md
  • DEVELOPMENT.md
• Description: Define explicit rules for deciding remove|keep|rewrite|defer, with hard guards for parity/contract tests and high-risk behavior surfaces.
• Dependencies:
  • Task 1.1
• Complexity: 3
• Complexity notes: Policy synthesis; complexity is in avoiding over-deletion.
• Acceptance criteria:
  • Rubric includes removal prerequisites and stop conditions.
  • Allowlist captures contract-critical tests/patterns that cannot be dropped without replacement.
  • Criteria are specific enough for deterministic triage.
• Validation:
  • test -s "$AGENT_HOME/out/workspace-test-cleanup/decision-rubric.md"
  • test -s "$AGENT_HOME/out/workspace-test-cleanup/contract-allowlist.tsv"
  • rg -n 'remove|keep|rewrite|defer|contract|parity|exit code|json' "$AGENT_HOME/out/workspace-test-cleanup/decision-rubric.md"

Task 1.4: Produce sprint execution manifest and crate-tier lanes

• Location:
  • $AGENT_HOME/out/workspace-test-cleanup/execution-manifest.md
  • $AGENT_HOME/out/workspace-test-cleanup/crate-tiers.tsv
  • docs/plans/workspace-obsolete-tests-cleanup-plan.md
• Description: Convert scan outputs into executable crate tiers (safe, medium, high-risk) and task lanes, with dependencies and explicit gating per tier.
• Dependencies:
  • Task 1.2
  • Task 1.3
• Complexity: 4
• Complexity notes: Dependency shaping across many crates with overlap constraints.
• Acceptance criteria:
  • All stale-test candidates are assigned to a tier and owning sprint task.
  • High-overlap crates are serialized explicitly.
  • Manifest is deterministic and directly executable.
• Validation:
  • test -s "$AGENT_HOME/out/workspace-test-cleanup/execution-manifest.md"
  • test -s "$AGENT_HOME/out/workspace-test-cleanup/crate-tiers.tsv"
  • awk -F '\t' 'NR>1 && ($2=="" || $3=="" || $4=="") {print $0}' "$AGENT_HOME/out/workspace-test-cleanup/crate-tiers.tsv" | wc -l | tr -d ' ' | rg '^0$'

[graysurf]

Sprint 1 Ready for Review

• Sprint: 1 (Inventory and stale-test decision graph)
• Tasks in sprint: 4
• Note: Main-agent requests sprint-level review before merge/acceptance on the plan issue (the issue remains open).
• PR values come from current Task Decomposition; unresolved tasks remain TBD until PRs are linked.

Task	Summary	PR
S1T1	Build workspace stale-test scanner and normalized output schema	#262
S1T2	Build unused-helper call graph for test scaffolding modules	#262
S1T3	Define stale-test decision rubric and contract-protection allowlist	#262
S1T4	Produce sprint execution manifest and crate-tier lanes	#262

Main-Agent Notes

Sprint 1 lane PR #262 submitted for main-agent review; awaiting CI + merge.

[graysurf]

Main-agent review approval for Sprint 1. PR #262 is approved for merge after CI success.

[graysurf]

Sprint 1 Accepted

• Sprint: 1 (Inventory and stale-test decision graph)
• Tasks in sprint: 4
• Note: Main-agent records sprint acceptance after merge gate passes and sprint rows are synced to done (issue remains open for remaining sprints).
• PR values come from current Task Decomposition; unresolved tasks remain TBD until PRs are linked.
• Approval comment URL: Plan: Workspace obsolete test cleanup and de-risking #261 (comment)

Task	Summary	PR
S1T1	Build workspace stale-test scanner and normalized output schema	#262
S1T2	Build unused-helper call graph for test scaffolding modules	#262
S1T3	Define stale-test decision rubric and contract-protection allowlist	#262
S1T4	Produce sprint execution manifest and crate-tier lanes	#262

Main-Agent Notes

Sprint 1 approved and merged via PR #262.

[graysurf]

Sprint 2 Start

• Sprint: 2 (Safe helper and dead-code-shield cleanup)
• Tasks in sprint: 5
• Note: Main-agent starts this sprint on the plan issue and dispatches implementation to subagents.
• Execution Mode comes from current Task Decomposition for each sprint task.

Task	Summary	Execution Mode
S2T1	Remove orphaned helper functions in low-risk crates	per-sprint
S2T2	Reduce obsolete #[allow(dead_code)] and crate-level dead-code shields in test helpers	per-sprint
S2T3	Consolidate nils-test-support helper primitives (atomic lane A)	per-sprint
S2T4	Adopt consolidated helpers in medium-risk crates (atomic lane B)	per-sprint
S2T5	Tier-2 regression sweep and fallout fixes	per-sprint

Sprint 2: Safe helper and dead-code-shield cleanup

Goal: Remove provably unused test helpers and unnecessary dead-code allowances in low/medium-risk crates.
PR grouping intent: per-sprint.
Execution Profile: parallel-x2 (parallel width 2).
Demo/Validation:

• Command(s):
  • cargo test -p nils-git-scope -p nils-image-processing -p nils-api-testing-core
  • cargo clippy --all-targets --all-features -- -D warnings
• Verify:
  • Helper deletions do not break tests.
  • Dead-code suppression is reduced only where safe.
    Sprint scorecard:
• TotalComplexity: 17
• CriticalPathComplexity: 13
• MaxBatchWidth: 2
• OverlapHotspots: crates/*/tests/common.rs, crates/*/tests/support/mod.rs, shared helper import boundaries in test modules.
  Parallelizable tasks:
• Task 2.1 and Task 2.2 can run in parallel after Task 1.4.
• Task 2.3 and Task 2.4 are split atomic phases before final regression sweep.

Task 2.1: Remove orphaned helper functions in low-risk crates

• Location:
  • crates/git-scope/tests/common.rs
  • crates/image-processing/tests/common.rs
  • crates/api-testing-core/tests/support/mod.rs
  • crates/git-summary/tests/common.rs
  • crates/plan-tooling/tests/common.rs
• Description: Remove helpers with zero callsites in low-risk crates and clean corresponding unused imports.
• Dependencies:
  • Task 1.4
• Complexity: 4
• Complexity notes: Mostly mechanical, with small risk of hidden transitive references.
• Acceptance criteria:
  • Orphaned helper symbols are removed.
  • No new warnings/errors in affected crate tests.
  • Behavior tests still pass unchanged.
• Validation:
  • cargo test -p nils-git-scope -p nils-image-processing -p nils-api-testing-core -p nils-git-summary -p nils-plan-tooling
  • before="$(awk -F '\t' 'NR>1 && $5=="allow_dead_code" {c++} END{print c+0}' "$AGENT_HOME/out/workspace-test-cleanup/stale-tests.tsv")"; after="$( (rg -n 'allow\\(dead_code\\)' crates/git-scope/tests/common.rs crates/image-processing/tests/common.rs crates/api-testing-core/tests/support/mod.rs crates/git-summary/tests/common.rs crates/plan-tooling/tests/common.rs || true) | wc -l | tr -d ' ' )"; test "$after" -le "$before"

Task 2.2: Reduce obsolete #[allow(dead_code)] and crate-level dead-code shields in test helpers

• Location:
  • crates/fzf-cli/tests/common.rs
  • crates/semantic-commit/tests/common.rs
  • crates/git-lock/tests/common.rs
  • crates/memo-cli/tests/support/mod.rs
  • crates/agent-docs/tests/common.rs
• Description: Remove or narrow dead-code allowances where helper usage now proves they are unnecessary; keep allowances only for intentionally shared helper APIs.
• Dependencies:
  • Task 1.4
• Complexity: 4
• Complexity notes: Moderate because different crates rely on different helper exposure patterns.
• Acceptance criteria:
  • Broad crate-level dead-code allowances are narrowed where possible.
  • Remaining allowances have clear necessity.
  • No clippy/regression failures are introduced.
• Validation:
  • cargo test -p nils-fzf-cli -p nils-semantic-commit -p nils-git-lock -p nils-memo-cli -p nils-agent-docs
  • cargo clippy -p nils-fzf-cli -p nils-semantic-commit -p nils-git-lock -p nils-memo-cli -p nils-agent-docs --tests -- -D warnings

Task 2.3: Consolidate nils-test-support helper primitives (atomic lane A)

• Location:
  • crates/nils-test-support/src/cmd.rs
  • crates/nils-test-support/src/git.rs
  • crates/nils-test-support/src/stubs.rs
  • crates/nils-test-support/src/fs.rs
• Description: Consolidate duplicated helper primitives in nils-test-support first, so crate-level helper cleanup can adopt a stable shared base.
• Dependencies:
  • Task 2.1
  • Task 2.2
• Complexity: 3
• Complexity notes: Focused shared-base changes reduce downstream churn.
• Acceptance criteria:
  • Shared helper primitives cover the duplicate wrapper patterns identified in Sprint 1.
  • Public helper behavior remains deterministic for existing tests.
  • No crate-level test helper is changed in this task.
• Validation:
  • cargo test -p nils-test-support
  • cargo clippy -p nils-test-support --tests -- -D warnings

Task 2.4: Adopt consolidated helpers in medium-risk crates (atomic lane B)

• Location:
  • crates/git-cli/tests/common.rs
  • crates/git-lock/tests/common.rs
  • crates/semantic-commit/tests/common.rs
  • crates/plan-issue-cli/tests/common.rs
  • crates/macos-agent/tests/common.rs
  • crates/macos-agent/tests/real_common.rs
• Description: Replace crate-local duplicate wrappers with shared primitives from Task 2.3 and remove obsolete wrapper layers.
• Dependencies:
  • Task 2.3
• Complexity: 2
• Complexity notes: Small adoption pass by design to keep this task atomic.
• Acceptance criteria:
  • Target helper modules compile and run without behavior regressions.
  • Duplicate wrapper surface is reduced in listed files.
  • Helper APIs used by tests remain clear and minimal.
• Validation:
  • cargo test -p nils-git-cli -p nils-git-lock -p nils-semantic-commit -p nils-plan-issue-cli -p nils-macos-agent

Task 2.5: Tier-2 regression sweep and fallout fixes

• Location:
  • crates/git-cli/tests/commit.rs
  • crates/plan-issue-cli/tests/task_spec_flow.rs
  • crates/macos-agent/tests/cli_smoke.rs
  • crates/screen-record/tests/recording_test_mode.rs
  • crates/git-summary/tests/edge_cases.rs
  • crates/fzf-cli/tests/edge_cases.rs
• Description: Run crate-targeted sweeps after helper cleanup and repair any behavior regressions introduced by removed stale helpers.
• Dependencies:
  • Task 2.4
• Complexity: 4
• Complexity notes: Breadth-driven task; complexity from diagnostics and focused fixes.
• Acceptance criteria:
  • All targeted crate test suites are green.
  • Any changed assertions remain contract-accurate.
  • No unresolved stale-helper references remain in targeted crates.
• Validation:
  • cargo test -p nils-git-cli -p nils-plan-issue-cli -p nils-macos-agent -p nils-screen-record -p nils-git-summary -p nils-fzf-cli
  • count="$(rg -n 'mod common;|mod support;' crates/git-cli/tests crates/plan-issue-cli/tests crates/macos-agent/tests crates/screen-record/tests crates/git-summary/tests crates/fzf-cli/tests | wc -l | tr -d ' ')"; test "$count" -gt 0

[graysurf]

Sprint 2 subagent lane update (S2T1..S2T5) is in progress on branch issue/s2-t1-remove-orphaned-helper-functions-in-low-risk-cra (worktree: issue-s2-t1). Draft PR: #263.

[graysurf]

Sprint 2 Ready for Review

• Sprint: 2 (Safe helper and dead-code-shield cleanup)
• Tasks in sprint: 5
• Note: Main-agent requests sprint-level review before merge/acceptance on the plan issue (the issue remains open).
• PR values come from current Task Decomposition; unresolved tasks remain TBD until PRs are linked.

Task	Summary	PR
S2T1	Remove orphaned helper functions in low-risk crates	#263
S2T2	Reduce obsolete #[allow(dead_code)] and crate-level dead-code shields in test helpers	#263
S2T3	Consolidate nils-test-support helper primitives (atomic lane A)	#263
S2T4	Adopt consolidated helpers in medium-risk crates (atomic lane B)	#263
S2T5	Tier-2 regression sweep and fallout fixes	#263

Main-Agent Notes

Sprint 2 lane PR #263 published by subagent; awaiting main-agent review and CI.

[graysurf]

Main-agent review approval for Sprint 2. PR #263 is approved for merge after CI success.

[graysurf]

Sprint 2 Accepted

• Sprint: 2 (Safe helper and dead-code-shield cleanup)
• Tasks in sprint: 5
• Note: Main-agent records sprint acceptance after merge gate passes and sprint rows are synced to done (issue remains open for remaining sprints).
• PR values come from current Task Decomposition; unresolved tasks remain TBD until PRs are linked.
• Approval comment URL: Plan: Workspace obsolete test cleanup and de-risking #261 (comment)

Task	Summary	PR
S2T1	Remove orphaned helper functions in low-risk crates	#263
S2T2	Reduce obsolete #[allow(dead_code)] and crate-level dead-code shields in test helpers	#263
S2T3	Consolidate nils-test-support helper primitives (atomic lane A)	#263
S2T4	Adopt consolidated helpers in medium-risk crates (atomic lane B)	#263
S2T5	Tier-2 regression sweep and fallout fixes	#263

Main-Agent Notes

Sprint 2 approved and merged via PR #263.

[graysurf]

Sprint 3 Start

• Sprint: 3 (High-risk obsolete behavior test cleanup)
• Tasks in sprint: 4
• Note: Main-agent starts this sprint on the plan issue and dispatches implementation to subagents.
• Execution Mode comes from current Task Decomposition for each sprint task.

Task	Summary	Execution Mode
S3T1	Identify deprecated/removed-path tests and fixture dependencies	per-sprint
S3T2	Remove obsolete deprecated-path tests and stale fixtures	per-sprint
S3T3	Rebaseline parity and contract suites for current command surfaces	per-sprint
S3T4	Full workspace regression gate after high-risk cleanup	per-sprint

Sprint 3: High-risk obsolete behavior test cleanup

Goal: Remove deprecated-path and no-longer-reachable behavior tests while preserving parity coverage.
PR grouping intent: per-sprint.
Execution Profile: serial (parallel width 1).
Demo/Validation:

• Command(s):
  • cargo test --workspace
  • ./.agents/skills/nils-cli-verify-required-checks/scripts/nils-cli-verify-required-checks.sh
• Verify:
  • Deprecated-path cleanup does not reduce contractual behavior coverage.
  • Any removed behavior tests are replaced by updated contract tests when needed.
    Sprint scorecard:
• TotalComplexity: 15
• CriticalPathComplexity: 15
• MaxBatchWidth: 1
• OverlapHotspots: parity/characterization suites (parity_oracle, parity_guardrails, CLI contract tests), fixture snapshots tied to legacy command surfaces.
  Parallelizable tasks:
• none (intentional serial execution due high contract-overlap risk).

Task 3.1: Identify deprecated/removed-path tests and fixture dependencies

• Location:
  • crates/codex-cli/tests/parity_oracle.rs
  • crates/gemini-cli/tests/parity_oracle.rs
  • crates/plan-issue-cli/tests/parity_guardrails.rs
  • crates/plan-issue-cli/tests/output_contract.rs
  • crates/api-test/tests/fixtures/grpc/smoke.suite.json
  • crates/plan-issue-cli/tests/fixtures/shell_parity/help.txt
  • $AGENT_HOME/out/workspace-test-cleanup/deprecated-path-candidates.tsv
• Description: Build a definitive list of tests that only guard deprecated or removed command surfaces, including fixture coupling and replacement needs.
• Dependencies:
  • Task 2.5
• Complexity: 3
• Complexity notes: Analysis-heavy task with moderate fixture-tracing complexity.
• Acceptance criteria:
  • Candidate file header is exactly: id\tcrate\tpath\ttest_or_fixture\trisk\treplacement\tdecision\trationale.
  • Candidate list has risk grade and replacement decision per entry.
  • Entries without replacement are explicitly justified as true dead coverage.
  • No contract-critical tests are misclassified as removable.
• Validation:
  • test -s "$AGENT_HOME/out/workspace-test-cleanup/deprecated-path-candidates.tsv"
  • awk -F '\t' 'NR>1 && ($5=="" || $6=="") {print $0}' "$AGENT_HOME/out/workspace-test-cleanup/deprecated-path-candidates.tsv" | wc -l | tr -d ' ' | rg '^0$'

Task 3.2: Remove obsolete deprecated-path tests and stale fixtures

• Location:
  • crates/codex-cli/tests/parity_oracle.rs
  • crates/gemini-cli/tests/parity_oracle.rs
  • crates/plan-issue-cli/tests/parity_guardrails.rs
  • crates/git-cli/tests/dispatcher.rs
  • crates/api-test/tests/fixtures/grpc/smoke.suite.json
  • crates/plan-issue-cli/tests/fixtures/shell_parity/help.txt
• Description: Delete tests and fixtures classified as obsolete in Task 3.1 where no reachable behavior remains.
• Dependencies:
  • Task 3.1
• Complexity: 5
• Complexity notes: Elevated risk due potential hidden contract coupling.
• Acceptance criteria:
  • Only approved obsolete tests/fixtures are removed.
  • Removed entries no longer appear in references.
  • Workspace tests continue passing after deletion batch.
• Validation:
  • cargo test --workspace
  • if [ -f "$AGENT_HOME/out/workspace-test-cleanup/deprecated-path-candidates.tsv" ]; then header="$(head -n1 "$AGENT_HOME/out/workspace-test-cleanup/deprecated-path-candidates.tsv")"; test "$header" = $'id\tcrate\tpath\ttest_or_fixture\trisk\treplacement\tdecision\trationale'; fi
  • if [ -f "$AGENT_HOME/out/workspace-test-cleanup/deprecated-path-candidates.tsv" ]; then awk -F '\t' 'NR>1 && $7=="remove" {print $3}' "$AGENT_HOME/out/workspace-test-cleanup/deprecated-path-candidates.tsv" | while read -r p; do [ -z "$p" ] || test ! -e "$p" || { echo "still exists: $p"; exit 1; }; done; fi

Task 3.3: Rebaseline parity and contract suites for current command surfaces

• Location:
  • crates/codex-cli/tests/parity_oracle.rs
  • crates/gemini-cli/tests/parity_oracle.rs
  • crates/plan-issue-cli/tests/parity_guardrails.rs
  • crates/plan-issue-cli/tests/cli_contract.rs
  • crates/plan-issue-cli/tests/output_contract.rs
  • crates/api-rest/tests/cli_smoke.rs
  • crates/api-gql/tests/cli_smoke.rs
• Description: Update parity/contract tests to reflect current supported surfaces after obsolete-path removal while preserving behavioral guarantees.
• Dependencies:
  • Task 3.2
• Complexity: 4
• Complexity notes: Contract sensitivity is high; changes must stay explicit and minimal.
• Acceptance criteria:
  • Parity suites target current command surfaces only.
  • Contract assertions still validate output/JSON/exit invariants.
  • No unexplained expectation drift.
• Validation:
  • cargo test -p nils-codex-cli -p nils-gemini-cli -p nils-plan-issue-cli -p nils-api-rest -p nils-api-gql
  • cargo test --workspace --tests

Task 3.4: Full workspace regression gate after high-risk cleanup

• Location:
  • DEVELOPMENT.md
  • .agents/skills/nils-cli-verify-required-checks/scripts/nils-cli-verify-required-checks.sh
• Description: Run full required checks and coverage gate to confirm high-risk stale-test cleanup is safe for workspace delivery.
• Dependencies:
  • Task 3.3
• Complexity: 3
• Complexity notes: Execution-heavy but operationally straightforward.
• Acceptance criteria:
  • Required checks entrypoint passes.
  • Coverage remains >= 85.00%.
  • Any failures have explicit remediation notes before merge.
• Validation:
  • ./.agents/skills/nils-cli-verify-required-checks/scripts/nils-cli-verify-required-checks.sh
  • cargo llvm-cov nextest --profile ci --workspace --lcov --output-path target/coverage/lcov.info --fail-under-lines 85
  • scripts/ci/coverage-summary.sh target/coverage/lcov.info