Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Shell helper scripts for the template research framework. Covers system health checks, local CI reproduction, uv bootstrap, and rsync backup tiers. Scripts live at scripts/ root level.
.claude/skills/docxology-template-shell/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-02 | ✗→✓ | ▲ Improved | 28% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -45% | 0% |
| case-05 | ✗→✓ | ▲ Improved | -32% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -42% | 0% |
Shell helper scripts for the template framework.
Load this skill when you need to:
run.sh| Script | Purpose | |--------|---------| | scripts/shell/health-check.sh | Pre-flight check (Python, uv, disk, Docker, repo) | | scripts/shell/ci_local.sh | Local CI reproduction | | scripts/shell/shell_bootstrap.sh | uv bootstrap + sandbox env vars | | scripts/shell/bash_utils.sh | Shared helpers (do not source directly in pipeline) | | scripts/shell/backup-daily.sh | Site-configured daily .hermes rsync tier | | scripts/shell/backup-weekly.sh | Site-configured weekly repository-cache rsync tier | | scripts/shell/backup-full.sh | Named, metadata-bearing .hermes/.cache/output snapshot | | scripts/shell/restore-test.sh | Private-scratch metadata and transfer-consistency verification |
shell_bootstrap.sh is sourced by run.sh — do not alter its export names.bash_utils.sh is for backup/health scripts and integration tests only.rsync, ssh, and an SSH alias oruser@hostname positional argument (default: backup); they do not use a BACKUP_DEST environment variable.
--local-root <absolute-dir> arguments for a disposable localfull-backup/restore round trip. Local success does not prove off-site availability, creation-time integrity, encryption, retention, or source quiescence.
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 10,628 | 33,732 | +217% | 1 | 1 | 0% | 1,743 | 1,112 | -36% | 0 | 0 | — |
case-02 | fail→pass | 5,988 | 4,262 | -29% | 1 | 1 | 0% | 903 | 1,153 | +28% | 0 | 0 | — |
case-03 | fail→pass | 14,403 | 5,997 | -58% | 1 | 1 | 0% | 2,627 | 1,445 | -45% | 0 | 0 | — |
case-04 | pass→pass | 11,066 | 3,224 | -71% | 1 | 1 | 0% | 1,829 | 1,028 | -44% | 0 | 0 | — |
case-05 | fail→pass | 9,978 | 3,210 | -68% | 1 | 1 | 0% | 1,487 | 1,011 | -32% | 0 | 0 | — |
case-06 | pass→pass | 44,956 | 3,785 | -92% | 1 | 1 | 0% | 2,265 | 1,081 | -52% | 0 | 0 | — |
case-07 | fail→pass | 12,047 | 4,129 | -66% | 1 | 1 | 0% | 1,871 | 1,080 | -42% | 0 | 0 | — |
case-08 | fail→pass | 11,883 | 1,935 | -84% | 1 | 1 | 0% | 1,778 | 782 | -56% | 0 | 0 | — |
case-09 | pass→pass | 13,318 | 5,093 | -62% | 1 | 1 | 0% | 1,933 | 1,188 | -39% | 0 | 0 | — |
case-10 | fail→pass | 8,040 | 2,175 | -73% | 1 | 1 | 0% | 1,324 | 783 | -41% | 0 | 0 | — |
case-11 | fail→pass | 7,848 | 1,752 | -78% | 1 | 1 | 0% | 1,112 | 681 | -39% | 0 | 0 | — |
case-12 | fail→pass | 15,320 | 2,305 | -85% | 1 | 1 | 0% | 2,220 | 819 | -63% | 0 | 0 | — |
case-13 | fail→pass | 11,097 | 4,022 | -64% | 1 | 1 | 0% | 1,704 | 987 | -42% | 0 | 0 | — |
case-14 | fail→pass | 5,547 | 1,595 | -71% | 1 | 1 | 0% | 894 | 671 | -25% | 0 | 0 | — |
case-15 | fail→pass | 11,694 | 3,137 | -73% | 1 | 1 | 0% | 1,726 | 904 | -48% | 0 | 0 | — |
case-16 | fail→pass | 9,708 | 2,672 | -72% | 1 | 1 | 0% | 1,398 | 877 | -37% | 0 | 0 | — |
case-17 | fail→pass | 6,990 | 2,431 | -65% | 1 | 1 | 0% | 1,023 | 776 | -24% | 0 | 0 | — |
case-18 | pass→pass | 11,582 | 1,771 | -85% | 1 | 1 | 0% | 1,748 | 673 | -61% | 0 | 0 | — |
case-19 | pass→pass | 14,388 | 5,338 | -63% | 1 | 1 | 0% | 2,414 | 1,297 | -46% | 0 | 0 | — |
case-20 | fail→pass | 6,618 | 6,138 | -7% | 1 | 1 | 0% | 1,120 | 1,503 | +34% | 0 | 0 | — |
case-21 | fail→pass | 8,950 | 4,924 | -45% | 1 | 1 | 0% | 1,578 | 1,296 | -18% | 0 | 0 | — |
case-22 | fail→pass | 10,206 | 5,611 | -45% | 1 | 1 | 0% | 1,586 | 1,339 | -16% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +77 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/4/2026 | +76% |
Other measured skills in the registry, with their headline benchmark lift.