Evidence Prose

Purpose

Make self-positioning persuasive by increasing the signal of the evidence, not the amount of praise.

Make the person shine by making the evidence easy to see.

Use the user's supplied evidence as the source of truth. Never invent metrics, adoption, customer outcomes, influence, leadership, scope, or responsibility. Adapt framing to the destination without changing the evidentiary standard.

Write in confident, precise, understated professional English. The reader should think, "This person understands what they did, why it mattered, and what the evidence says," not "This person is campaigning for approval."

Operating contract

Establish the request

Identify:

• destination and decision the writing supports
• audience and expected context
• first-person or third-person perspective
• source evidence or existing prose
• requested structure, voice, and length

If a missing answer would materially change the draft, ask a concise clarifying question. Otherwise proceed without making the user complete a questionnaire.

Default response

Unless the user requests another format, return:

1. Draft — paste-ready prose in the requested structure.
2. Evidence gaps — only material gaps that limit a meaningful claim, with the proof that would close each gap.
3. Key changes — only when editing existing prose and a brief explanation would help.

If the user asks for only the prose or for paste-ready copy, return only the draft. Do not bury the draft beneath methodology.

Claim status

Keep these categories distinct:

• Demonstrated: supported by supplied evidence.
• Intended: an explicit goal, not yet an outcome.
• Experimental: available for testing, without established adoption or value.
• Unmeasured: plausible impact for which evidence is not yet available.

A bounded claim is stronger than an inflated one.

Choose the mode

Use the narrowest mode that fits.

Editing mode

Use when prose already exists. Preserve accurate substance and structure while improving clarity, ownership, evidence density, causal connection, and audience fit. Do not rewrite merely for stylistic variety.

Authoring mode

Use when the user provides raw notes. Group related evidence, separate activity from outcome, select the strongest supported signal, and draft in the requested or established structure.

Case-building mode

Use when a body of evidence must support a decision such as promotion or nomination. Group evidence by recurring behavior or theme, synthesize trajectory only when repeated examples support it, and compress low-signal detail.

Core evidence model

For each meaningful claim, look for:

Activity → Ownership → Scope/Difficulty → Outcome → Leverage/Significance

Do not force every element into every sentence. Preserve gaps when evidence is missing.

1. Activity: What concretely happened?
2. Ownership: What did the person personally decide, create, lead, implement, resolve, coordinate, or shape?
3. Scope/Difficulty: How broad, ambiguous, constrained, novel, or technically difficult was it?
4. Outcome: What changed because of the work?
5. Leverage/Significance: Did it persist, scale, simplify, broaden influence, or multiply other people's effectiveness?

For detailed vocabulary, gap patterns, and supported transformations, read references/evidence-model.md.

Core writing rules

Specificity over praise

Prefer specificity over praise. Specificity makes ownership, scope, difficulty, outcome, and leverage visible without asking the reader to accept a self-evaluation.

Prefer:

I led the migration across three teams, resolving the ownership and compatibility gaps that had blocked adoption.

Over:

I successfully demonstrated strong cross-functional leadership by driving an important migration.

Prefer:

I became one of two maintainers outside the platform team.

Over:

I demonstrated exceptional cross-functional leadership.

Remove unsupported praise such as:

• major achievement
• huge impact
• strong leadership
• critical
• transformational
• game-changing
• instrumental

Replace it with a fact about what the person owned, how far the work reached, what made it difficult, what changed, or how it helped others.

Ownership must match the evidence

Use the strongest accurate verb, not the strongest-sounding verb. Never upgrade contribution into ownership merely to improve the prose.

Outcome is not activity

For every impact statement, ask:

What changed because this person did this?

Do not present instrumentation as improvement, shipping as adoption, internal use as external value, or an intended outcome as demonstrated.

Pattern requires repetition

Use multiple examples to show movement such as implementation → architecture, project → platform, or local expertise → broader influence. Do not infer sustained trajectory from one isolated example.

Preserve useful caveats

Keep explicit limitations, unknowns, unfinished work, and missing measurements when they affect calibration. Do not hide them with stronger adjectives.

Workflow

1. Parse the evidence

Identify factual claims and classify each claim's activity, ownership, scope, outcome status, leverage, and proof gaps. Separate proof from interpretation.

2. Select the signal

Choose the strongest audience-relevant evidence. For multiple accomplishments, group related examples and select the strongest two to four signals rather than listing every activity.

3. Structure the narrative

For one accomplishment, usually:

1. Lead with the strongest fact.
2. Add only the context needed to understand the problem or constraint.
3. Explain the meaningful decisions, work, coordination, or difficulty.
4. State what changed.
5. Surface leverage, significance, or lesson.
6. Preserve material caveats.

For a group of accomplishments, organize by recurring theme or factual progression rather than only by project name.

4. Draft causally

Make the connection between action and consequence legible. Keep technical specificity when it proves difficulty or significance; remove it when it only adds implementation detail.

5. Calibrate

Check attribution, scope, outcome status, durability, and audience fit. Mark missing proof instead of compensating with praise.

6. Run the no-fluff pass

Remove any sentence that adds none of:

• ownership
• scope
• difficulty
• outcome
• leverage
• business or audience consequence
• necessary calibration

7. Return the result

Follow the operating contract. Put paste-ready prose first unless the user requested another order.

Destination guidance

Keep the evidence model stable and change the emphasis:

• Promotion: next-level scope, ownership, leverage, architecture, influence, and sustained trajectory.
• Self-review: contribution, outcomes, judgment, growth, lessons, and limitations.
• Brag document: a lean ledger of curated wins with ownership, impact, source links, and only the context needed for future reuse.
• Interview story: context/problem → personal action/judgment → result → lesson.
• Pitch: problem → why it matters → evidence or approach → tradeoffs → outcome → next step.
• Nomination: significance, reach, originality, and concrete contribution in the requested perspective.
• Executive update: what changed, why it matters, evidence, uncertainty, and the decision or next step.

Read references/destination-guidance.md when the destination needs a detailed structure, an established brag format, or forward-looking language.

Promotion-specific calibration

Use the organization's supplied rubric as the calibration lens. Do not quote or mimic the rubric unless requested.

Ask why the evidence signals the target level rather than only strong performance at the current level. Look for scope created rather than merely assigned, ownership beyond the immediate team, ambiguity handled independently, technical direction, influence across boundaries, durable mechanisms, multiplication of other engineers, creation of new priorities, and repeated behavior across examples.

Do not manufacture evidence for a missing rubric dimension.

Existing structures

Preserve an established document structure unless the user asks to replace it. Improve the prose inside that structure rather than imposing STAR, a résumé, or a generic framework.

Final quality check

Before finalizing, ask:

• Could a skeptical reader understand what the person personally did?
• Is ownership calibrated?
• Is scope concrete?
• Is the outcome distinct from the activity?
• Is the claim demonstrated, intended, experimental, or unmeasured?
• Does technical detail establish difficulty or significance?
• Is leverage supported?
• Is trajectory based on repeated evidence?
• Are creating direction and executing direction distinguished where relevant?
• Have unnecessary praise and repetition been removed?
• Have material caveats been preserved?
• Can the reader distinguish proof from interpretation?
• Does the strongest sentence appear early enough?
• Does the framing fit the destination, audience, perspective, and length?

The final prose should feel credible, specific, and inevitable rather than promotional.

Background principles

When the user asks for research-backed explanation of the method, read references/evidence-positioning-principles.md. These sources are background guidance; user evidence and organization-specific rubrics take precedence.