# Agent Skill Bundle — Conversation-to-Source

> Generated from the canonical public Skill source. This bundle is an invocation artifact, not a separate editable source.

## Agent instruction

Apply this Skill to the user’s actual task. Follow its inputs, behavior, guardrails, outputs, and verification criteria. Do not merely explain or summarize the Skill. If required task context is missing, ask only for the missing information that blocks correct execution.

## Skill identity

- source: skills/conversation-to-source
- canonical human page: https://wbeen-personal-kb.vercel.app/skills/conversation-to-source/
- version: 0.1.0
- workflow phase: execute
- invocation: either

## Skill contract

Turn a long AI conversation into distinct public-ready source candidates while preserving sensitivity boundaries and review decisions.

### Inputs

- Conversation export or working notes
- Intended audience and publishing boundary
- Candidate asset types to look for

### Outputs

- One candidate source per reusable asset
- Sensitivity and review classification
- Evidence pointer back to the source material
- Unresolved items requiring human review

### Guardrails

- Generalize or omit private, client, company, and personal details.
- Classify ambiguity as review rather than silently publishing it.
- Preserve enough provenance to re-check the extracted candidate.
- Never treat extraction as publication approval.

### Verification

- Every candidate has a declared sensitivity state.
- A reviewer can trace the candidate to its source location.
- Private or ambiguous material is excluded from public output.

### Source instructions

## What it does

Conversation-to-Source separates reusable public-source candidates from the private conversation that produced them. It treats extraction as a classification and review step, not as automatic publishing.

## When to reach for it

Use it after a long AI-assisted working session when decisions, insights, workflows, or templates are useful beyond the original context and need a safer source boundary.

## What one run does

1. Identify reusable assets and describe their intended audience.
2. Create one candidate source per asset instead of copying the transcript.
3. Attach sensitivity and review status to every candidate.
4. Preserve a source pointer and unresolved questions.
5. Hand candidates to a scoped review or publishing step.

## Common failure modes

- Treating a coherent conversation as automatically public.
- Copying names, internal systems, credentials, or personal details into a candidate.
- Losing provenance so a reviewer cannot re-check an extraction.
- Publishing the candidate before a human boundary check.

## It’s working if

A reviewer can understand each candidate without the private transcript, see its sensitivity classification, and trace it back to evidence when needed.

## Where it fits

This is an **Execute**-phase extraction skill. It sits before Scope & Acceptance Criteria and Public-safe Publishing, which decide whether a candidate is ready for a specific public outcome.

## Completion rule

Return the useful task result first. Then state briefly which verification criteria passed, which could not be checked, and any stop/escalation condition that remains.