Starter kits

test-coverage-loop — the 100%-coverage loop

Adds tests until the suite reaches full coverage, on demand. Each beat drafts a test for the largest uncovered region in a throwaway worktree and…

Adds tests until the suite reaches full coverage, on demand. Each beat drafts a test for the largest uncovered region in a throwaway worktree and reports whether it raised coverage without weakening what "covered" means. It never commits a test itself.

Category: J · Testing & QA · Heartbeat: conditional (run-until-done, on-demand — no schedule) · Level: L1 (report-only, drafts-not-commits) · Cost: High

Source: Loop #5, "100% Test Coverage Loop," from Forward Future's Loop Library, original prompt by Matthew Berman: "Add tests until we have 100% test coverage."

Quickstart

1. Read LOOP.md — the six parts, the prompt, the limits, the three stops,
   and the source's own caveat about coverage vs. assertion quality.
2. test-coverage-loop-state.md is your spine — already committed with an
   empty per-region register.
3. Claude Code: run it in-session on demand
   ("/loop Run the test-coverage-loop skill until 100%..."), pointed at
   .claude/skills/loop-task/SKILL.md. OpenCode: rename
   opencode.json.example → opencode.json, add the AGENTS.md line, run
   skills/loop-task.md via a capped for-loop.
4. Rehearse once, by hand, on real uncovered code before running to
   completion: "Run the test-coverage-loop skill now for one beat."
5. Grade the rehearsal with the loop-verifier agent — it re-runs the suite
   and checks whether the drafted test actually asserts anything. PASS →
   run it to completion. FAIL → fix the skill, never force a hollow test
   through for the coverage number.

What's in this kit

File	Role
`LOOP.md`	the definition — six parts, prompt, limits, ownership, three stops
`test-coverage-loop-state.md`	the spine — per-region coverage status, retry counts, the `exclusion-candidate` register
`loop-budget.md`	caps: 40 runs, 400k tokens, 5 regions/beat, 1 retry/region — added by this library; the source specified none
`loop-constraints.md`	the constitution — never commits a test, never drafts a hollow assertion
`loop-run-log.md`	append-only, one line per beat
`.claude/skills/loop-task/SKILL.md`	the procedure (Claude Code)
`.claude/agents/loop-verifier.md`	the read-only reviewer — re-runs the suite and checks assertion quality, not just the coverage delta
`opencode.json.example` · `skills/loop-task.md`	the OpenCode twin — same skill text, capped run

Loop Ready checklist

Provable success condition — the source's own spec, verbatim: "the full test suite passes at 100% coverage"
Run limit — 40 runs, 5 regions/beat, 1 retry/region (added by this library)
Spine written first, committed empty
Report-only (L1) start — drafts tests in a throwaway worktree, commits nothing
Human gate placed — you review each drafted test's assertions, not just the percentage
One log line per beat, no silent runs
Kill switch — loop-pause-all in loop-budget.md

How this relates to the course

This kit adapts Forward Future's Loop #5 almost unmodified — its stopping condition ("the full test suite passes at 100% coverage") and checker ("the project's coverage report as the source of truth") are used verbatim, since they're already exactly the kind of machine-checkable spec this course's Step 5/B move demands. What this library adds: the L1-first shipping rule (kit-state.md; CLAUDE.md rule 4), the worktree isolation from Step 8, and the retry-once-then-exclusion-candidate doom-loop bound from infinite-loops.md, scenario 1. The source's own implementation note — coverage tools measure execution, not assertion quality — is carried directly into the loop-verifier's rubric and the human gate.

Full attribution: resources/sources.md.