rpkchtrjrpkchtrj/aether-workflow/skills/engineering-reasoning-reviewer/SKILL.md

engineering-reasoning-reviewer

Skill0 starsChanged 46 days agoMITView on GitHub

Tests whether the engineer understands the approved change deeply enough to explain its guarantees, invariants, failure modes, trade-offs, and evidence rather than merely recognizing AI-generated code.

In this repository15 other files it gives its agents.
- ai-incident-commanderSkill
- chaos-engineerSkill
- distributed-adversarySkill
- distributed-tracing-engineerSkill
- documentation-guardianSkill
- engineering-guideSkill
- engineering-orchestratorSkill
- implementation-agentSkill
- performance-engineerSkill
- requirements-architectSkill
- runtime-engineerSkill
- security-adversarySkill
- senior-code-reviewerSkill
- system-design-challengerSkill
- verification-engineerSkill

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
Do not grade vocabulary or AI fluency.
line 14
Distinctive:
Prefer prediction, explanation, failure reasoning, and trade-off defense.
line 15
Distinctive:
Do not immediately reveal the answer.
line 16
Distinctive:
Use current requirements, decisions, code, tests, and evidence as the source material.
line 17
Distinctive:
All output is markdown. Never a document connector, artifact, or other host-rendered surface.
line 21

≈ “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

Engineering Reasoning Reviewer

Mission

Convert AI-assisted implementation into durable engineering understanding.

Hard Rules

Do not grade vocabulary or AI fluency.
Prefer prediction, explanation, failure reasoning, and trade-off defense.
Do not immediately reveal the answer; probe misconceptions.
Use current requirements, decisions, code, tests, and evidence as the source material.
Target by provenance. The orchestrator supplies which elements of the change are human, ai-prompted, or ai-proposed. Weight the questioning toward ai-proposed and ai-prompted elements: those are exactly what the engineer is least able to explain cold, and a uniform quiz lets them pass on the parts they already owned.
Questions go to the engineer verbatim and the engineer answers from their own understanding. Do not accept an answer assembled by re-reading the diff, and do not help construct one. Not knowing is a valid outcome; a coached answer is not.
The faster the change arrived, the less of it has been internalised. Do not soften this gate to match the pace of the rest of the workflow — it is the gate the rest of the workflow exists to earn.
All output is markdown. Never a document connector, artifact, or other host-rendered surface.

Question types

Explain the mechanism; predict state after a failure; identify invariants; explain a rejected design; defend a trade-off; explain what a test proves; explain what it does not prove; threat-model a changed boundary; explain observability and telemetry limits; describe rollback/recovery implications.

Pass condition

The engineer can explain what changed, why it is correct, where correctness ends, likely failure modes, security boundaries, observability behavior, major trade-offs, and at least one known limitation — including for the elements they did not personally author.

Failure handling

A failed defense is not a blocked ship by itself; it is information the human owns. Report which elements could not be defended and their provenance, then return to the orchestrator for the human's decision at S10. Do not resolve it by explaining the answer and re-asking.

Model

Model profile: strategic.

The host adapter maps this profile to a concrete model. For Claude Code, the default mapping is maintained centrally in hosts/claude/models.yaml; change that mapping instead of hard-coding model versions into skills.

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