Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →This skill should be used when the user asks to "generate Vault configurations", "plan secret rotation", "analyze vault audit logs", "manage secrets lifecycle", or "set up HashiCorp Vault".
.claude/skills/borghei-secrets-vault-manager/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-03 | ✗→✓ | ▲ Improved | 670% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -18% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 6% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 12% | 0% |
| case-07 | ✗→✓ | ▲ Improved | 9% | 0% |
> Category: Engineering > Domain: Secrets Management & Security
The Secrets Vault Manager skill provides tools for generating HashiCorp Vault configurations, planning and scheduling secret rotation cycles, and analyzing vault audit logs for suspicious access patterns. Essential for teams managing secrets at scale.
Before generating configs or plans, confirm these inputs. If any is unknown or vague, ASK — do not assume:
--env, --secrets-engines; shapes the generated HCL)Stop rule: ask only the 2-3 that most change the output. If the user says "just draft it," proceed and list your assumptions at the top of the artifact.
bash# Generate Vault configuration python scripts/vault_config_generator.py --env production --secrets-engines kv,database,transit # Plan secret rotation schedule python scripts/rotation_planner.py --inventory secrets_inventory.json # Analyze vault audit logs python scripts/audit_log_analyzer.py --log-file vault_audit.log --format json
| Tool | Purpose | Key Flags | |------|---------|-----------| | vault_config_generator.py | Generate HashiCorp Vault configurations | --env, --secrets-engines, --auth-methods | | rotation_planner.py | Plan and schedule secret rotation cycles | --inventory, --policy, --format | | audit_log_analyzer.py | Analyze vault audit logs for anomalies | --log-file, --time-range, --format |
vault_config_generator.py to generate HCL configsrotation_planner.py to generate scheduleaudit_log_analyzer.py for anomaly detection| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→fail | 14,613 | 16,720 | +14% | 1 | 1 | 0% | 2,441 | 3,480 | +43% | 0 | 0 | — |
case-01 | fail→fail | 22,103 | 20,856 | -6% | 1 | 1 | 0% | 3,939 | 4,549 | +15% | 0 | 0 | — |
case-03 | fail→pass | 3,453 | 17,237 | +399% | 1 | 1 | 0% | 482 | 3,713 | +670% | 0 | 0 | — |
case-04 | fail→pass | 9,972 | 5,022 | -50% | 1 | 1 | 0% | 1,731 | 1,425 | -18% | 0 | 0 | — |
case-05 | fail→pass | 7,752 | 3,575 | -54% | 1 | 1 | 0% | 1,163 | 1,237 | +6% | 0 | 0 | — |
case-06 | fail→pass | 7,572 | 3,190 | -58% | 1 | 1 | 0% | 1,036 | 1,158 | +12% | 0 | 0 | — |
case-07 | fail→pass | 5,874 | 2,209 | -62% | 1 | 1 | 0% | 911 | 990 | +9% | 0 | 0 | — |
case-08 | pass→pass | 15,234 | 6,148 | -60% | 1 | 1 | 0% | 2,311 | 1,698 | -27% | 0 | 0 | — |
case-09 | fail→pass | 19,706 | 4,385 | -78% | 1 | 1 | 0% | 2,751 | 1,272 | -54% | 0 | 0 | — |
case-10 | pass→pass | 12,480 | 4,434 | -64% | 1 | 1 | 0% | 1,787 | 1,336 | -25% | 0 | 0 | — |
case-11 | fail→pass | 10,570 | 5,100 | -52% | 1 | 1 | 0% | 1,557 | 1,409 | -10% | 0 | 0 | — |
case-12 | fail→fail | 18,467 | 12,974 | -30% | 1 | 1 | 0% | 3,111 | 2,677 | -14% | 0 | 0 | — |
case-13 | pass→pass | 18,682 | 18,143 | -3% | 1 | 1 | 0% | 2,790 | 3,466 | +24% | 0 | 0 | — |
case-14 | fail→pass | 19,248 | 17,869 | -7% | 1 | 1 | 0% | 2,975 | 3,418 | +15% | 0 | 0 | — |
case-15 | fail→pass | 15,142 | 11,301 | -25% | 1 | 1 | 0% | 2,501 | 2,493 | -0% | 0 | 0 | — |
case-16 | fail→pass | 10,655 | 6,035 | -43% | 1 | 1 | 0% | 1,880 | 1,691 | -10% | 0 | 0 | — |
case-17 | pass→pass | 7,162 | 3,392 | -53% | 1 | 1 | 0% | 1,032 | 1,176 | +14% | 0 | 0 | — |
case-18 | fail→pass | 6,624 | 3,318 | -50% | 1 | 1 | 0% | 907 | 1,139 | +26% | 0 | 0 | — |
case-19 | fail→pass | 11,829 | 6,873 | -42% | 1 | 1 | 0% | 1,653 | 1,662 | +1% | 0 | 0 | — |
case-20 | pass→pass | 9,683 | 12,904 | +33% | 1 | 1 | 0% | 1,615 | 3,168 | +96% | 0 | 0 | — |
case-21 | pass→pass | 17,598 | 14,735 | -16% | 1 | 1 | 0% | 3,233 | 3,262 | +1% | 0 | 0 | — |
case-22 | pass→pass | 17,841 | 17,816 | -0% | 1 | 1 | 0% | 2,911 | 3,408 | +17% | 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 +55 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.
Other measured skills in the registry, with their headline benchmark lift.