Reasoning
Analysis

Find the decision rationales, tradeoffs, and operating patterns in expert conversations and source material. Turn those explanations into structured inputs for AI Agent Skills.

Make the logic behind the action visible.

  1. 01 Recover the rationale
  2. 02 Compare the context
  3. 03 Develop the skill logic

01 / Recover the rationale

Recover the structure inside an explanation

An expert can compress years of experience into a few sentences. Their explanation may move quickly from an observed signal to an assumption, a rejected alternative, and a decision.

Reasoning Analysis helps identify those connections across sessions and source material. Examine why an action made sense, which conditions shaped it, and where the explanation leaves a question for further investigation.

Start with the expert conversation
Rationale extraction

Connect the observation to the choice it informed.

Perspective alignment

Compare the assumptions behind different recommendations.

Source context linking

Keep the explanation connected to material a reviewer can inspect.

02 / Unpack a decision

What made the expert choose that path?

Select a part of this illustrative release-recovery explanation. Follow its source passage into the rationale and the evidence that still needs examination.

Illustrative expert explanation
“The error rate rose after deployment. The earlier version still depended on the field the migration removed. Rolling back could restore the code while leaving the data incompatible. We held the rollback and asked the application owner to confirm a compatible recovery path.”

The highlighted passage stays attached to its interpretation.

SignalConstraintTradeoffAction
An illustrative structure for examining a decision. Each part remains connected to the expert’s explanation.
What drew attention

Start with what the expert observed

The explanation starts with a symptom and its timing. Capture that observation so the connection to the deployment can be investigated.

Evidence to examine

Compare the reported timing with the deployment record and relevant system observations.

What limited the choice

Expose the condition behind the decision

The earlier version depends on a field the migration removed. That dependency changes which recovery paths deserve consideration.

Evidence to examine

Connect the application version, migration notes, and expert explanation. Confirm that they refer to the same environment.

What the alternative could cost

Keep the rejected option in the explanation

Restoring the code may appear to offer a quick recovery. The expert weighs that benefit against an incompatible data model and explains why the usual action could fail.

Evidence to examine

Identify the assumption behind the proposed benefit and the dependency behind the risk. Preserve both in the developing rationale.

What happened next

Keep the action and its boundary together

The expert holds the rollback and seeks confirmation from the application owner. The explanation establishes a need for further evidence before choosing a recovery path.

Evidence to examine

Confirm who can resolve the compatibility question and which evidence would support the next action.

03 / Compare perspectives

Compare the conditions behind different advice

Different recommendations may reflect different constraints. Compare each explanation with its source context to understand what differs and where further expert review is needed.

Recover the setting

Identify the system, version, objective, and circumstances each expert was discussing.

Find the assumption that changes the answer

Trace the recommendation back to the dependency or operating condition it relies on.

Keep unresolved disagreement visible

When accounts still conflict under the same conditions, preserve both rationales and identify what evidence could resolve the difference.

Explore the research workflow
Illustrative comparison

Two recommendations, two operating conditions

Expert AExpert BContextReview
Bring the accounts into a shared context for comparison. The reviewer examines whether the difference is explained.
Expert A considers a rollback
The earlier application version and the data model remain compatible.
Expert B holds the rollback
The migration has removed a field the earlier application version expects.
Question for the reviewer
Does the evidence support these different conditions, and which one applies to the skill’s intended use?

04 / Develop skill inputs

Give the next stage an explanation it can inspect

Bring the rationale, supporting passages, assumptions, and open questions into the developing skill. Keep candidate rules available for review and testing as the evidence grows.

01 / Evidence

A rationale with its sources

Connect each interpretation to the session excerpt and supporting material. Review whether the evidence supports the decision and its operating conditions.

Explore AI Skill Validation
02 / Boundaries

A condition to exercise

Turn the assumptions and exceptions into questions for scenario testing. Examine how the skill responds when the circumstances change.

Explore the AI Skill Playground
03 / Open questions

A gap with a next step

Make missing explanations and unresolved differences visible in the review process. Use coverage and readiness information to focus further expert input.

Explore Skill Lifecycle Analytics

Start with the explanation behind the work

Which expert decision deserves a closer look?

Bring a workflow, an expert conversation, and the material behind it. Explore the conditions, tradeoffs, and judgment that could become part of an AI Agent Skill.

RapidKT Plain Text Website