ariaxhanariaxhan/kernel-claude/skills/tearitapart/SKILL.md

tearitapart

Skill15 starsChanged 4 months agoMITView on GitHub

Critical pre-implementation review. Find what AI breaks. Verdict: PROCEED, REVISE, or RETHINK. Triggers: review plan, tear apart, critique, analyze.

In this repository30 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- llms.txtllms.txt
- app-devSkill
- architectureSkill
- buildSkill
- checkpointSkill
- context-mgmtSkill
- debugSkill
- dreamSkill
- evalSkill
- experimentSkill
- forgeSkill
- frontendSkill
- governance-syncSkill
- handoffSkill
- helpSkill
- human-passSkill
- ingestSkill
- initSkill
- knowledge-graphSkill
- landing-pageSkill
- marketing-siteSkill
- metricsSkill
- orchestrationSkill
- qualitySkill
- retrospectiveSkill
- reviewSkill
- shipSkill
- simplifySkill

The file

We split this skill into the rules it gives an agent

8 rules. About 0 are common, 6 are distinctive, and 2 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
Check AgentDB for prior contracts
line 35
Very specific:
Read _meta/research/ for anti-patterns
line 36
Very specific:
Run Big 5 checks from skills/quality/SKILL.md:
line 45
Distinctive:
Use quick_checks from quality skill for detection.
line 53
Distinctive:
No error path testing </phase>
line 83
Distinctive:
No Big 5 violations. Security passes.
line 98
Distinctive:
Use AskUserQuestion when: verdict is REVISE or RETHINK Ask: "Verdict: {REVISE|RETHINK}.
line 113
Distinctive:
Ask: "Verdict: {REVISE|RETHINK}.
line 114

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.

Files

SKILL.md

<skill id="tearitapart">

<purpose>
Pre-implementation review. Check Big 5, security, testing, architecture.
Goal: find real problems, not generic concerns.
</purpose>

<skill_load> Load: skills/quality/SKILL.md, skills/build/reference/testing.md, skills/tearitapart/reference/security.md Reference: skills/quality/reference/quality-research.md </skill_load>

<on_start>

agentdb read-start

</on_start>

<phase id="1_gather">
- Read plan/spec
- List files to be touched
- Check git status
- Check AgentDB for prior contracts
- Read _meta/research/ for anti-patterns

output: scope: N files tier: 1|2|3 prior_work: contracts, research found </phase>

<phase id="2_big5">
Run Big 5 checks from skills/quality/SKILL.md:

input_validation: Zod schema? Parameterized queries?
edge_cases: null, empty, unicode, timeout?
error_handling: no empty catch? Logged with context?
duplication: same logic repeated?
complexity: functions < 30 lines?

Use quick_checks from quality skill for detection. </phase>

<phase id="3_security">
Load: skills/tearitapart/reference/security.md

critical:

No hardcoded secrets
Auth tokens in httpOnly cookies
Rate limiting
HTTPS enforced

injection:

SQL: parameterized only
XSS: DOMPurify
CSRF: tokens on state changes</phase>

<phase id="4_testing">
Load: skills/build/reference/testing.md

verify:

Tests exist BEFORE implementation?
Edge cases covered?
Assertions specific (not toBeTruthy)?
Mocks at boundaries only?

red_flags:

"Will add tests later"
100% coverage, weak assertions
No error path testing</phase>

<phase id="5_architecture">
Load: skills/architecture/SKILL.md

verify:

Follows existing patterns?
Interface stability?
Modular boundaries?
Dependency direction correct?</phase>

<verdict>
<PROCEED>
No Big 5 violations. Security passes. Tests defined.
Output: "PROCEED with caveats: [list]"
</PROCEED>

<REVISE>
1-2 Big 5 violations (fixable). Missing security items.
Output: "REVISE: [changes with file:line]"
</REVISE>

<RETHINK>
3+ Big 5 violations. Fundamental security gaps. No tests.
Output: "RETHINK: [why flawed] → [alternative]"
</RETHINK>

<ask_user> Use AskUserQuestion when: verdict is REVISE or RETHINK Ask: "Verdict: {REVISE|RETHINK}. Want details on specific findings, or proceed with fixes?" Options: show details, proceed with fixes, override and proceed anyway </ask_user> </verdict>

<output_format> Save to _meta/reviews/{feature}-teardown.md:

# Tear Down: {feature}
reviewed: {timestamp}
tier: {1|2|3}
scope: {N files}

## Big 5
input_validation: pass|fail
edge_cases: pass|fail
error_handling: pass|fail
duplication: pass|fail
complexity: pass|fail

## Verdict: PROCEED | REVISE | RETHINK
{reasoning}

## Action Items
1. {fix with file:line}

</output_format>

<on_complete>

agentdb write-end '{"command":"tearitapart","verdict":"X","big5_violations":N}'

</on_complete>

</skill>

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free