Complete Test Coverage for Remaining Untested Scripts

[github-actions]

Strategic Improvement

Write test suites for the 3 remaining untested scripts to achieve >90% test coverage target. This closes a quality gap and ensures all platform components have regression protection. Scripts: sw-tmux-role-color.sh (81 lines), sw-tmux-status.sh (151 lines), sw-tracker-github.sh (requires implementation).

Acceptance Criteria

• Test suite scripts/sw-tmux-role-color-test.sh with PASS/FAIL counting
• Test suite scripts/sw-tmux-status-test.sh with PASS/FAIL counting
• Test suite scripts/sw-tracker-github-test.sh (or mark as not-implemented if provider incomplete)
• All test suites registered in package.json test script
• Mock tmux environment for pane color and status tests
• Test coverage metric updated to >90%

Context

• Priority: P0
• Complexity: fast
• Generated by: Strategic Intelligence Agent
• Strategy alignment: P0: Reliability & Success Rate (Technical Excellence)

[github-actions]

⚓ Shipwright Pipeline Starting

Field	Value
Template	standard
Complexity	standard
Max Iterations	20
Triage	Closes quality gap by adding regression protection for 3 core untested scripts (tmux utilities, tracker provider); straightforward test harness work following established patterns.
Runner	ubuntu-latest
Risk Level	unknown
Workflow	24270076437

The autonomous pipeline is processing this issue. Updates will be posted as stages complete.

[github-actions]

[github-actions]

[github-actions]

📋 Implementation Plan

Click to expand full plan (7 lines)

The plan covers creating 3 test suites for the remaining untested scripts:

1. sw-tmux-role-color-test.sh — Tests all 8 role→color mappings plus edge cases (unknown role, case-insensitive matching, tmux unavailable)
2. sw-tmux-status-test.sh — Tests stage_color, stage_icon, pipeline_widget, agent_widget, and dispatch routing
3. sw-tracker-github-test.sh — Tests all provider_* functions with mock gh binary, including NO_GITHUB guard and error cases

All follow the established test harness pattern using lib/test-helpers.sh, and get registered in package.json.

---

Generated by shipwright pipeline at 2026-04-11T00:27:21Z

[github-actions]

[github-actions]

📐 Architecture Decision Record

Click to expand ADR (477 lines)

Design: Complete Test Coverage for Remaining Untested Scripts

Context

Problem: The Shipwright project has 102 test suites covering 85+ scripts, but three utility scripts remain untested:

• scripts/sw-tmux-role-color.sh (81 lines) — Maps agent roles to tmux pane border colors
• scripts/sw-tmux-status.sh (151 lines) — Renders pipeline status widgets (stage colors, icons, agent states)
• scripts/sw-tracker-github.sh (TBD) — GitHub provider adapter wrapping the gh CLI

Constraints:

• Shell script testing (Bash 3.2 compatible, no associative arrays/readarray)
• External dependency mocking (tmux, gh CLI) required for isolation
• Test harness established in lib/test-helpers.sh with PASS/FAIL counting, mock binaries in /tmp, colored output
• All tests registered in package.json and run via npm test
• These scripts are in the critical path: tmux integration and GitHub provider are used by daemon, pipeline orchestration, and fleet operations

Why This Matters: These three scripts handle critical infrastructure (tmux pane management, status display, GitHub API routing). Untested edge cases here can cascade into CI/CD failures and silent fallback behavior.

---

Decision

Create 3 dedicated test suites following the established pattern, each fully isolating external dependencies via mock binaries and temporary directories. Each suite validates:

1. Happy path: Expected behavior with valid inputs
2. Error cases: Missing dependencies, invalid arguments, API failures
3. Edge cases: Empty/unusual inputs, race conditions, fallback behavior

Key Architectural Choices:

Choice	Rationale
One test file per script	Mirrors 102 existing test suites; makes coverage intent clear in package.json
Mock binaries in /tmp	Isolates external deps (tmux, gh); enables offline testing; prevents side effects
Reuse lib/test-helpers.sh	Leverage established PASS/FAIL counting, error trapping, colored output
NO_GITHUB guard testing	These scripts respect $NO_GITHUB env var; must validate fallback behavior
Shell-native assertions	No external test runners (Jest, Vitest) for bash; keep test logic close to the code under test

---

Alternatives Considered

Alternative	Pros	Cons	Decision
Skip testing (maintain 102 suite count)	Reduces scope, faster shipping	Leaves critical infrastructure untested; violates project's "test everything" policy; future bugs harder to catch	✗ Rejected
Merge into existing suites (e.g., add tmux tests to sw-tmux-test.sh)	Less file overhead	Existing test suites are already 200+ lines; lib/test-helpers.sh is shared utilities, not test suite; modularity requires one suite per script	✗ Rejected
Use different test framework (Bats, shUnit2)	More expressive assertions	Breaks consistency with 102 existing suites; adds dependency; requires retraining on new patterns	✗ Rejected
Use Vitest for shell mocking (via mock-shell or jest-shell)	Unified test runner	Adds JS/TypeScript dependency for bash testing; mocking gets opaque (hard to debug mock failures); established pattern already works	✗ Rejected
Inline tests in the scripts themselves	Self-documenting code	No separation of concerns; scripts become bloated; test logic leaks into production	✗ Rejected

Chosen Approach: Create 3 focused test suites in scripts/, register in package.json, use established mock pattern and lib/test-helpers.sh.

---

Implementation Plan

Files to Create

Path	Lines	Purpose	Dependencies

---

Generated by shipwright pipeline design stage at 2026-04-11T00:29:36Z

[github-actions]

[github-actions]

🔨 Build started — shipwright loop with 20 max iterations, 1 agent(s), model: haiku

[github-actions]

[github-actions]

[github-actions]