# Bullet Coaching

Use this for bullet diagnosis, rewriting, ordering, restraint, and approval.
Read `impact-excavation.md` only when important evidence is missing.

## Contents

1. Editorial test
2. Visible coaching loop
3. Specificity and translation
4. Construction
5. Weak patterns
6. Portfolio and final QA

## 1. Editorial test

Test every candidate bullet:

1. **Purpose:** What must the target reader believe?
2. **Ownership:** What did Darsh personally decide, build, lead, recommend, or
   change?
3. **Result:** What decision or consequence followed?
4. **Maturity:** Does it show judgment, scale, difficulty, or tradeoffs?
5. **Restraint:** Is it ornate, self-important, over-engineered, or try-hard?
6. **Technical calibration:** Does each implementation detail help this reader?
7. **Specificity relevance:** Does each proper noun, segment, metric, and
   domain term change the reader's conclusion?
8. **Clarity:** Can a recruiter outside the company understand it in one read?
9. **Industry recognition:** Is the impact recognized in the source function
   and valued by the target reader?
10. **Portfolio value:** Does it add a distinct proof point?
11. **Compression:** Can it fit within two rendered lines?
12. **Truth:** Are the fact, number, causality, title, and ownership supported?

Reject or rewrite a bullet that fails truth or clarity. Do not average a
high-stakes failure into an otherwise strong score.

After the truth check, run a competitive-strength check:

- Is this the strongest credible verb?
- Are the commercial stakes prominent?
- Did process language bury the decision or result?
- Is the strategic implication supported and interview-defensible?

Accuracy is not modesty. Prefer the bolder defensible version.

## 2. Visible coaching loop

Work one bullet at a time:

1. Display the current bullet and its role in the resume story.
2. Diagnose the main weakness.
3. Ask one targeted question when important evidence is missing.
4. Rewrite after the missing material is known.
5. Present alternatives only when they encode a real tradeoff.
6. Ask the user to lock the wording.
7. On explicit approval, persist it using `persistence-and-state.md`.

Do not silently edit a locked bullet. Reopen it and obtain new approval.

## 3. Specificity and translation

Require a specificity dividend. A detail must improve target fit, clarify the
business model or decision, prove scale or difficulty, distinguish the
candidate, or increase credibility.

Ask:

- Would the target recruiter understand it without insider context?
- Does it explain the transferable mechanism or merely name its source-domain
  version?
- Would removing it weaken ownership, scale, result, or target fit?
- Does it reinforce a screening stereotype the resume is trying to overcome?

Use:

```text
internal term -> understandable domain term -> transferable business mechanism
```

Choose the lowest level that earns its explanation cost. Run a deletion test:
if removing the detail leaves the same conclusion, cut or translate it.

## 4. Construction

Use:

```text
Strong verb + work or decision + scale or context + outcome
```

Aim for 25-35 words and two rendered lines. Lead with the most
decision-relevant point, not necessarily chronology or the largest number.
Keep one idea per bullet.

Use verbs that encode actual ownership:

- `built` for an artifact
- `led` for coordinated execution
- `owned` for sustained accountability
- `recommended` for a consequential decision
- `redesigned` for a changed system

Do not rotate verbs merely for variety. Use numbers to prove scale,
consequence, or change. Use keywords only when they accurately name the work.

## 5. Weak patterns

- **Activity without outcome:** ask what changed.
- **Vague quantification:** ask for baseline, new state, period, and consumer.
- **Adjacency language:** replace `supported` only when stronger ownership is
  true.
- **Stitched bullets:** split or cut the weaker idea.
- **Insider jargon:** translate it.
- **Weak verb:** name the artifact, recommendation, decision, or system.
- **Process win:** connect speed or accuracy to capacity, revenue, risk, or
  operating cadence.
- **Tool-name dropping:** subordinate the tool to the artifact and consequence.
- **Implementation dumping:** retain mechanics only when selection-relevant.
- **Inflated sophistication:** replace abstractions with the actual work.
- **Try-hard affinity:** show credible usage or work, not praise or destiny.
- **Strategy without action:** name the recommendation, tradeoff, or lever.

## 6. Portfolio and final QA

Balance the resume across strategic judgment, commercial outcomes, operating
systems, customer insight, cross-functional ownership, and analytical fluency.
Do not let every bullet prove the same capability.

Within a role, lead with the most target-relevant achievement and use another
strong metric-backed bullet as the closer.

Before assembly:

- check tense, titles, dates, and verb repetition
- remove generic filler and try-hard language
- verify every bullet advances the accepted story
- verify technical terms serve the reader
- identify the three messages retained after a 20-second scan