Create Personal Identity

This skill is an action router — pick the step that matches the user's intent and execute only that step. Do not run other steps; do not parallelize.

Personal identity owns how an author sounds and reasons, never employer positioning, product claims, or corporate editorial policy.

Before any action, read both contracts completely:

skills/create-personal-identity/references/lifecycle.md
skills/create-personal-identity/references/package-contract.md

For Step 1 or Step 2, also read the shared storage contract before writing:

skills/blog-writer/references/identity-storage.md

That contract lives with blog-writer because every identity type shares its destination, discovery, and selection layout.

Use this approval loop for every consequential create or update:

• Set status to draft.
• Show the one-paragraph summary, Required and Avoid guidance, high-impact inferences, and unresolved items.
• Ask one focused question at a time for consequential conflicts.
• Apply corrections, re-read each changed conclusion beside its cited sources, and present the corrected conclusion with those sources.
• Set status to approved only after explicit approval.

Step 1 — Create a Personal Identity

1. Inventory every supplied file, directory, URL, pasted passage, existing skill, transcript, published post, interview, feedback item, and legacy persona file.

2. Classify each source as authored, ghostwritten, heavily corporate-edited, experimental, rejected, or unresolved; record its scope and authority before drawing conclusions.

3. Compile evidence about register, confidence, reader relationship, argument construction, rhetorical devices, humor, references, technical depth, recurring elements, deliberate variation, off-voice moves, and durable bio facts. Keep explicit preferences separate from observed patterns; sample frequency is evidence, not a quota.

4. Classify representative prose by writing mode and authorship. Route supported contiguous passages through an examples resource.

5. Mark a writing mode Unresolved when it lacks representative prose. Do not use these as prose evidence:

   • URLs whose prose has not been read
   • summaries
   • isolated lines

6. Resolve the destination through the shared storage contract, write the package through skills/create-personal-identity/references/package-contract.md, run the approval loop, then finish here.

Before approval, confirm every declared resource and consequential guidance item traces to an inventoried source; leave any missing trace Unresolved.

Step 2 — Migrate a Legacy Personal Identity

Read the source directory and run the exact v0-to-v1 migration in the lifecycle reference. Stop unless the directory matches that reference's legacy v0 shape. Resolve the destination through the shared storage contract. Run the approval loop and finish here.

Step 3 — Update a Personal Identity

Read the existing v1 package and its provenance before reading new evidence. Preserve valid guidance and source history. Identify which conclusions the new sources confirm, weaken, replace, or leave unresolved; never regenerate the package as if prior decisions did not exist. Stop on a non-v1 package; Step 2 owns the only supported older shape.

Apply accepted changes through skills/create-personal-identity/references/package-contract.md. Keep the existing status only when changes do not alter consequential guidance. For consequential changes, run the approval loop. Finish here.

Step 4 — Audit a Personal Identity

Compare the package and its provenance against the representative sources the user names. Report stale, unsupported, contradictory, overfit, and missing guidance with the evidence for each finding. Pay special attention to employer language misclassified as personal voice and observed patterns promoted into requirements.

Do not modify the identity during an audit. If the user later asks to apply accepted findings, that follow-up request routes to Step 3. Finish here.