test: consolidated test coverage across 9 modules (133 new tests)

[anandgupta42]

What does this PR do?

Consolidates 9 test PRs (#387, #390, #394, #395, #396, #397, #398, #399, #401) into one, discarding 4 duplicates (#391, #392, #393, #400). Adds 133 new tests across 13 modules and fixes 5 production bugs in finops-formatting.ts discovered during testing.

New test coverage:

• id.test.ts — 14 tests: Identifier generation, monotonic ordering, timestamp extraction, Zod schema
• shell.test.ts — 9 tests: shell blacklist enforcement for fish/nu, cache reset, fallback
• path-traversal.test.ts — 7 tests (new): Protected.isSensitiveWrite for .pem/.key, .gnupg, .env variants
• ignore.test.ts — 7 tests (new): FileIgnore.match directory patterns, globs, whitelist overrides
• paths-parsetext.test.ts — 16 tests: ConfigPaths.parseText env/file substitution, JSONC parsing
• finops-formatting.test.ts — 14 tests: formatBytes/truncateQuery edge cases
• finops-role-access.test.ts — 10 tests: RBAC formatGrants/formatHierarchy/formatUserRoles
• tool-lookup.test.ts — 4 tests: Zod schema introspection for tool parameters
• summary-git-path.test.ts — 10 tests: unquoteGitPath decoding for non-ASCII filenames
• tracing.test.ts — 2 tests (added): Recap loop detection boundaries
• patch.test.ts — 9 tests (added): maybeParseApplyPatchVerified entry point coverage
• instruction.test.ts — 13 tests (new): InstructionPrompt.loaded() dedup guards
• message-v2.test.ts — 23 tests (new): MessageV2.filterCompacted() boundary detection

Bug fixes (discovered during testing):

• finops-formatting.ts: fix formatBytes crash on negative/NaN/fractional inputs
• finops-formatting.ts: fix truncateQuery returning empty for whitespace-only, exceeding maxLen when < 4
• training-import.test.ts: fix pre-existing typecheck errors (missing context/rule mock keys)

Code review fixes:

• summary-git-path.test.ts: use tmpdir() instead of projectRoot to prevent test artifact leakage (Gemini finding)
• finops-role-access.test.ts: consolidate duplicate afterAll hooks for proper cleanup (GLM-5 finding)

Type of change

• New feature (non-breaking change which adds functionality)
• Bug fix (non-breaking change which fixes an issue)

Issue for this PR

Closes #402

How did you verify your code works?

259 tests, 584 assertions, 0 failures across 14 test files
Typecheck: PASS
Marker check: PASS (no upstream-shared files modified)

Multi-model code review by Claude, GPT 5.2 Codex, Gemini 3.1 Pro, and GLM-5 (4/6 models). All approve.

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• New and existing tests pass locally with my changes
• I ran linters/typechecks before committing

🤖 Generated with Claude Code

[claude]

Claude Code Review

This repository is configured for manual code reviews. Comment @claude review to trigger a review.

_{Tip: disable this comment in your organization's Code Review settings.}