On-demandOn-demand: invoked explicitly — never occupies your agent's context until called.
◆Coding pack
Last verified 8/16/2026
What it changes
Passed 14 of 19 test cases with the skill, and 10 of 19 without it.
Without skill55%
With skill73%
Checks it passed that the model alone missed
Triggers /create-pr to open the pull request
Focuses agent A on Architecture and MVVM compliance
Delegates PII scanning to the detect-data-leaks skill
and 2 more in the full report
5 fixed11 passed either way5 still failing1 broke
Pass-rate lift
+18 pp
Tokens
-22%
Agent turns
0%
Response time
-37%
Model gemini-3.6-flashMeasured 8/16/2026v1 · latest is v216 of 22 cases passed with skill · 19 comparable
Tokens, turns and time describe resource use, not a dollar price.
This benchmark measured an earlier version of the skill.
A newer attempt on gemini-3.6-flash on 10/2/2026 did not produce a publishable comparison (not comparable). The snapshot above is the previously measured run.
See it in action
One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 5
Prompt
After completing selected fixes from a pre-review audit, what next workflow command or skill should be triggered to publish the pull request?
What was checked, without and with the skill
Checked
Without skill
With skill
Triggers /create-pr to open the pull request
Missed
Passed
case-19 · gemini-3.6-flash · 8/16/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md27 linessingle file
Pre-Review
Deep code review
Process
Run git diff develop...HEAD to get the full diff.
Launch 5 code-reviewer agents in parallel, each with a different focus:
A — Architecture & MVVM compliance
B — Correctness (edge cases, null handling, error states)
C — Code quality & conventions (Lumen UI, TypeScript, naming, new deps). Reject new .pnpmfile.cjs hooks and pnpm.packageExtensions (skill pnpm-resolution).
D — General review — DRY, KISS, missing tests, anything that would slow down a reviewer or cause future regressions
E — Security: invoke the detect-data-leaks skill for PII; check for auth bypass, injection, SSRF, crypto misuse; scan the diff for hardcoded secrets, tokens, API keys, or private keys.
Each agent prompt: "Review the diff vs develop. Focus on [FOCUS]. Return findings with file:line, severity (🔴 Critical / 🟡 Suggestion / 🟢 Nice to have), and suggested fix."
Merge + deduplicate findings. Present grouped by severity:
🔴 Critical — path/to/file.ts:42 — issue → fix
🟡 Suggestion — ...
🟢 Nice to have — ...
Summary: X critical, Y suggestions, Z nice-to-haves
Ask the user which items to fix: "all criticals", "all", "#1,#2, #3", or "skip" → /create-pr