Ready to supercharge your AI agents?

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

Adversarial Verify · archive228 | DecimalAI

code reviewby archive228

Adversarial Verify

archive228/adversarial-verify

Source archive228/loopkit ↗License MIT

Review a diff against the goal spec assuming the code is BROKEN. The reviewer that lives in the maker's head always agrees with itself — this pulls review into a hostile, separate pass. Invoke after every code change before marking work done.

SkillScore73Maintained

Automatic

QA & Test Engineering pack

Last verified 8/16/2026

What it changes

Passed 16 of 22 test cases with the skill, and 6 of 22 without it.

Without skill27%

With skill73%

Checks it passed that the model alone missed

The shortcut value is 'scope creep'.
The shortcut is named 'pass-by-mock'.
The shortcut category is 'silent decision'.

and 21 more in the full report

16 fixed2 passed either way1 still failing3 broke

Pass-rate lift: +45 pp
Tokens: -15%
Agent turns: 0%
Response time: -35%

Model gemini-3.6-flashMeasured 8/16/2026v118 of 22 cases passed with skill

Tokens, turns and time describe resource use, not a dollar price.

Use this skill

Save it into your project's skills folder. Claude Code picks it up in its next session.

curl -fsSL --create-dirs -o .claude/skills/archive228-adversarial-verify/SKILL.md https://app.decimal.ai/s/archive228-adversarial-verify/SKILL.md

Want it pinned and updatable? pip install "decimalai>=0.15", then decimalai skills add archive228-adversarial-verify, and start Claude Code with claude --plugin-dir "$(git rev-parse --show-toplevel)/.agents"

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

pip install "decimalai>=0.15"

2 · Get the skill — writes .agents/skills/archive228-adversarial-verify/ and pins it in .decimal/skills.lock

decimalai skills add archive228-adversarial-verify

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

curl -fsSL --create-dirs -o .claude/skills/archive228-adversarial-verify/SKILL.md https://app.decimal.ai/s/archive228-adversarial-verify/SKILL.md

Raw file: /s/archive228-adversarial-verify/SKILL.md ↗ · machine-readable summary at /s/archive228-adversarial-verify.json ↗

Safety checks: 3 of 3 passed

Code scan CleanIntent PassedContent Clean

See it in action

One of 16 test cases the skill fixed. Same model, same prompt, with and without the skill.

1 of 16

Prompt

Can you audit the following pull request changes against the task requirements? Act as a hostile reviewer searching for fake fixes or hidden shortcuts that mask underlying bugs. Do not execute the code or provide corrected code. Return the evaluation purely as JSON containing a pass/fail status field and a list of specific line failures with their cause.

What was checked, without and with the skill
Checked	Without skill	With skill
The JSON output uses 'passes' as the boolean key instead of 'status'.	Missed	Passed
The JSON output uses 'failures' as the array key instead of 'line_failures'.	Missed	Passed
The JSON failure items use 'why' as the explanation key instead of 'cause'.	Missed	Passed

case-03 · 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.md31 linessingle file

Adversarial Verify

Default stance: the code is broken until proven otherwise. Your job is to find where. Do not be polite. Do not propose fixes. Do not run the code. Just hunt.

Read first

The goal spec (PROMPT.md / the task). What does "done" actually require?
The diff. Every changed line.

The 11 shortcuts agents take to fake "done" — check each

Relaxed tests — assertions weakened or deleted to make red go green.
Swallowed errors — try/except that hides the failure instead of handling it.
Fake renames — a function "fixed" by renaming, behavior unchanged.
Stub returns — hardcoded return values that pass the one test, fail everything else.
Comment-as-fix — the bug is now a TODO.
Happy-path only — 500s, empty inputs, missing files unhandled.
Scope creep — changes unrelated to the goal ("while I was in there").
Invented API — a method/param that doesn't exist in the actual source.
Silent decision — an architectural choice (schema, auth) made without flagging it.
Pass-by-mock — the test mocks the exact thing it claims to verify.
Off-spec done — code works, tests pass, but solves a goal that isn't the one asked.

Output (JSON, no prose)

{"passes": false, "failures": [{"line": 42, "shortcut": "swallowed errors", "why": "..."}]}

If it genuinely passes, say so in one line. Most of the time, it doesn't.

Track record

Benchmark

16 / 22

test cases passed, vs 6 / 22 without

Adoption

Users: 1
Times used: 11

Safety and source details

SkillSafety

✓ 3 of 3 passed

Code scan15 of 15 checks clean

0 critical · 0 warning · 0 info — scanner v4 — 8/3/2026

View 15 scanner checks

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

archive228 / loopkit

skills/adversarial-verify/SKILL.md

View on GitHub ↗

License MITImported 8/3/2026Version v1

More skills from the same source and nearby in the registry, with their headline benchmark lift.

Ready to supercharge your AI agents?

See it in action

What's inside

Adversarial Verify

Read first

The 11 shortcuts agents take to fake "done" — check each

Output (JSON, no prose)

Track record

Benchmark

Adoption

SkillSafety

Source

Related skills