Ready to supercharge your AI agents?

Install any skill in seconds. Free to start, no credit card required.

Get Started Free →

QC Gate · macro-inc | DecimalAI

macro-inc/qc

NewSkillScoreMaintained

Quality gate. 5 parallel agents review changes. All must pass.

Code Reviewby macro-inc

Links this skill to your agents and keeps it current with the author’s updates.

Automatic

QA & Test Engineering pack

1 · One-time — install the CLI (docs ↗)

pip install decimalai

2 · Get the skill — writes .claude/skills/macro-inc-qc/SKILL.md

decimalai skills pull macro-inc-qc --out .claude/skills/

No CLI? Get the same file — curl · raw file

curl -fsSL --create-dirs -o .claude/skills/macro-inc-qc/SKILL.md https://app.decimal.ai/s/macro-inc-qc/SKILL.md

Raw file: /s/macro-inc-qc/SKILL.md ↗ · machine-readable summary at /s/macro-inc-qc.json ↗

In productionGlobal

how this skill behaves on real traffic · last 30 days

No production data for this skill yet — usage numbers appear here once agents start running it.

Anonymized aggregate across every org using this skill — proof from real production use, not a demo.

Does this skill help? · Benchmark: with vs. withoutsign in to run this on your agents

23 test cases from the publisher’s eval.yaml · each run twice on gemini-3.6-flash — with the skill injected, then without · last run 8h ago

—-17% tokens0% turns74% of whole cases pass with skill (17/23)

Test case	Without → With	Effect	Δ tokens	Δ turns
case-05	✗→✓	▲ Improved	-2%	0%
case-07	✗→✓	▲ Improved	1%	0%
case-09	✗→✓	▲ Improved	-39%	0%
case-12	✗→✓	▲ Improved	11%	0%
case-13	✗→✓	▲ Improved	-1%	0%

Showing 5 of 23 cases

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v5 — 8/31/2026

live_secretremote_code_executionreverse_shelldata_exfiltrationssrf_metadataagent_snoopinganti_refusaldestructive_commandsobfuscated_unicodeunicode_homoglyphopaque_blobsuspicious_urltool_overreachtrigger_abuseprompt_injection_phrasing

IntentSafe

LLM judge · does the skill do what it claims

ContentClean

LLM judge · policy & harmful-content review

Source

Imported

macro-inc / macro

.agents/skills/qc/SKILL.md

View on GitHub ↗

License AGPL-3.0Imported 8/31/2026Version v1

Skill contents

SKILL.md

QC Gate

Setup

bash
DIFF=$(git diff main)
FILES=$(git diff main --name-only)
TASK="{task description or 'general changes'}"

Read 1-2 unmodified files from same directories for pattern context.

Agents

Spawn all 5 in parallel. All output raw JSON only, no markdown.

1: Code Review (sonnet)

DIFF: ${DIFF}

- Descriptive naming?
- Errors caught with useful messages?
- No hardcoded values, commented code, debug statements?
- No TODO without ticket ref?
- No obvious bugs?
- No useless comments?

{"pass": bool, "issues": [...]}

2: Simplification (sonnet)

DIFF: ${DIFF}

- Is this overcomplicated? Can I solve the same problem in a simpler manner?
- Can I reduce indirection?
- Can I reduce surface area?
- Premature abstraction? Premature Optimization? YAGNI violations?
- Dead code?
- Three similar lines > one abstraction

{"pass": bool, "issues": [...]}

3: Consistency (sonnet)

DIFF: ${DIFF}
PATTERNS: ${PATTERN_FILES}

- Matches existing codebase patterns?
- Proper types, no any, no unsafe casts?
- Idiomatic error handling?
- Changes internally consistent?
- Is logic isolated and composable?
- Are there existing tests if neccesary?

{"pass": bool, "issues": [...]}

4: Robustness (sonnet)

TASK: ${TASK}
DIFF: ${DIFF}

- Actually solves the problem?
- Edge cases: empty, null, zero, negative, boundaries, concurrency?
- Regression risk: changed signatures, shared state, removed exports?
- Maintains API contracts?

{"pass": bool, "issues": [...]}

5: Scope (haiku)

TASK: ${TASK}
FILES: ${FILES}
DIFF: ${DIFF}

- Solved the problem or just the symptom?
- Changes unrelated to task?
- Unnecessary refactoring?

{"pass": bool, "issues": [...]}

Results

| Check | Verdict | Issues | |-------|---------|--------| | Code Review | ✓/✗ | ... | | Simplification | ✓/✗ | ... | | Consistency | ✓/✗ | ... | | Robustness | ✓/✗ | ... | | Scope | ✓/✗ | ... |

ALL PASS: Ready to commit.

ANY FAIL: List issues, fix them, run /qc again.

Same issue 3x: Escalate to user.

Notes

Invalid JSON? Retry once. Still broken? Mark fail.
Large diffs (>500 lines): run simplification file-by-file.

Illustrative Want to see what macro-inc/qc would put in front of your model? Preview the rendered body and check its trigger phrases — no LLM call, no signup.

Latest benchmark · publisher runVerified

9/28/2026, 03:46 UTC · 23 cases

Measured on 20 comparable cases — below the 22 needed to publish a lift, so this result is shown here but not on the registry card or leaderboard.

Lift

+35%

74% vs 39%

Agent turns

0%

1.0 vs 1.0

Tokens

-17%

1,304.75 vs 1,569.8

fail → pass · 8pass → pass · 9fail → fail · 6

Per-case results (23)

Case	Status	Duration (ms)			Turns			Tokens			Tool calls
Case	Status	Without	With	Δ	Without	With	Δ	Without	With	Δ	Without	With	Δ
case-01 I just finished refactoring the payment gateway integration on my branch and want to run a complete quality control check against main before committing. The task description is 'update payment error handling and retry logic'. Please inspect the diff against main, check for overall code quality, code simplicity, pattern consistency against surrounding files, robustness against edge cases, and proper task scope, then present the results in a structured verdict table with issues listed.	fail→fail	4,019	6,323	+57%	1	1	0%	568	861	+52%	0	0	—
case-02 Please run a thorough quality gate audit on my local changes compared to main for the task 'add bulk export endpoint'. Evaluate the diff across code review standards, structural complexity, consistency with neighboring files in the codebase, edge case handling, and scope adherence. Format the final output as a checklist matrix showing pass/fail status for each dimension alongside any found issues.	fail→fail	8,252	6,441	-22%	1	1	0%	1,233	1,075	-13%	0	0	—
case-03 I'm preparing to commit my changes for 'fix user session timeout bug'. Compare my current branch diff against main, examine existing codebase patterns in neighboring files for context, and evaluate the changes for potential bugs, unnecessary abstractions, pattern violations, edge cases, and scope creep. Return a summary table detailing the outcome for each quality check along with next-step recommendations.	fail→fail	12,762	9,357	-27%	1	1	0%	1,766	843	-52%	0	0	—
case-04 I am preparing a pre-commit verification script for our feature branch. I usually compare branches using `git diff main...HEAD` or `git diff HEAD~1`. What exact git command should be used to extract the complete diff payload against main for quality gate evaluation?	fail→fail	14,625	12,156	-17%	1	1	0%	2,334	2,541	+9%	0	0	—
case-05 I need to inspect which files changed on my local feature branch relative to the main branch to determine evaluation scope. What command outputs just the names of changed files against main?	fail→pass	8,522	3,576	-58%	1	1	0%	1,157	1,136	-2%	0	0	—
case-06 When preparing pattern context before running consistency checks on modified files in `src/services/`, how many unmodified files from the same directory should be read to establish baseline patterns?	fail→fail	9,615	2,726	-72%	1	1	0%	1,347	869	-35%	0	0	—
case-07 We are building an automated quality gate runner that evaluates code review, simplification, consistency, robustness, and scope. Should these five checks be run sequentially in series or launched simultaneously in parallel?	fail→pass	14,858	131,973	+788%	1	1	0%	2,376	2,392	+1%	0	0	—
case-08 We want to assign cost-effective LLM models to our quality gate sub-agents. For the Scope check agent that evaluates task relevance and scope creep, which model tier should be selected among Sonnet and Haiku?	pass→pass	10,929	4,463	-59%	1	1	0%	1,576	1,227	-22%	0	0	—
case-09 Which LLM model tier is specified for executing the Code Review, Simplification, Consistency, and Robustness check agents in the quality gate process?	fail→pass	10,048	2,283	-77%	1	1	0%	1,376	834	-39%	0	0	—
case-10 When receiving responses from the five parallel review sub-agents, what format constraint should be placed on their output to ensure programmatic parsing?	pass→pass	13,063	4,356	-67%	1	1	0%	2,181	1,206	-45%	0	0	—
case-11 I have a large git diff containing 750 lines of changes across multiple modules. Does the quality gate process evaluate the entire diff at once for all checks, or is there a specific threshold where simplification analysis switches to a per-file strategy?	pass→pass	17,215	4,318	-75%	1	1	0%	2,584	1,267	-51%	0	0	—
case-12 All 5 quality checks (Code Review, Simplification, Consistency, Robustness, Scope) returned passing verdicts for my pull request diff. What exact status recommendation should be presented at the bottom of the report?	fail→pass	6,018	2,539	-58%	1	1	0%	884	977	+11%	0	0	—
case-13 During a quality gate run, the Robustness check failed due to missing boundary checks on zero values, while the other 4 checks passed. What workflow steps should the developer follow after fixing the code?	fail→pass	9,060	4,788	-47%	1	1	0%	1,356	1,345	-1%	0	0	—
case-14 A developer has run the quality gate check on their branch, but the Consistency check continues to fail for the exact same pattern violation. How many consecutive failures on the same issue trigger escalation to the user?	pass→pass	6,553	2,438	-63%	1	1	0%	986	912	-8%	0	0	—
case-15 When the Simplification sub-agent returns invalid JSON during execution, should the system immediately mark the check as failed, or is there a retry protocol?	pass→pass	12,555	2,993	-76%	1	1	0%	1,664	960	-42%	0	0	—
case-16 I am designing a custom reporter for our pre-commit audit gate. I was planning to use columns named `Category`, `Status`, and `Details`. What are the standard three column headers required for the results table?	fail→pass	13,212	3,686	-72%	1	1	0%	2,093	1,141	-45%	0	0	—
case-17 List all 5 quality check categories that must appear in the audit summary table after evaluating branch changes against main.	fail→fail	10,302	2,702	-74%	1	1	0%	1,337	939	-30%	0	0	—
case-18 What exact verdict symbols are used in the quality report table to denote passing and failing checks?	fail→pass	7,811	2,313	-70%	1	1	0%	986	899	-9%	0	0	—
case-19 What specific variables must be passed as inputs to the Scope review agent when evaluating whether a commit contains off-task changes?	pass→pass	13,499	3,587	-73%	1	1	0%	2,037	1,118	-45%	0	0	—
case-20 Unlike the general Code Review agent, the Consistency check agent requires extra contextual data. What parameter must be provided alongside the diff?	fail→pass	10,703	3,695	-65%	1	1	0%	1,428	1,066	-25%	0	0	—
case-21 I need to set up a GitHub Actions workflow in `.github/workflows/test.yml` to run `npm test` automatically whenever code is pushed to the main branch. Provide the YAML configuration for this CI pipeline.	pass→pass	6,150	5,041	-18%	1	1	0%	899	1,426	+59%	0	0	—
case-22 Write a Jest unit test suite for a function `parseConfig(filePath)` that verifies it returns an object when given a valid file path and throws an Error when given an empty path. Provide the test file code.	pass→pass	8,111	10,947	+35%	1	1	0%	1,338	2,218	+66%	0	0	—
case-23 We want to enforce code style in our repository using ESLint. Show me how to configure `.eslintrc.json` to enforce strict single quotes and require semi-colons.	pass→pass	8,339	7,717	-7%	1	1	0%	1,457	1,622	+11%	0	0	—

These cases were also screened by the skill’s intent / safety pass.See the Trust review ↗Reproduce this benchmark in Colab ↗

How this skill was measured

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. 23 cases were attempted, and 20 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +35 percentage points is the difference between those two pass rates over the 20 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.

Model measured: gemini-3.6-flash
Graded by: gemini-3.6-flash
Grading method: judged
Verification: verified
Pass rate without the skill: 39.1%
Pass rate with the skill: 73.9%
Cases attempted: 23
Cases counted toward lift: 20
Cases that regressed: 1
Skill version measured: v1
Run completed: 9/28/2026, 03:46 UTC

Measured example — same prompt, with vs. without the skill · from the benchmark run · case case-05

Prompt

I need to inspect which files changed on my local feature branch relative to the main branch to determine evaluation scope. What command outputs just the names of changed files against main?

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.

Show both answers

SkillScore blends 4 signalsno signals yet

BenchmarkRun a verified benchmark with positive lift

Live evalNeeds ≥10 evaluated traces in 30 days

AI ratingNeeds ≥10 rated traces

AdoptionNeeds ≥10 serves across ≥2 orgs in 30 days

No signals yet — run a benchmark or collect evaluated traces to earn a score.

Verified on gemini-3.6-flash · re-tested today

Other measured skills in the registry, with their headline benchmark lift.