impact-bullets

Impact Bullets Skill

Use this when

Use this skill to create or revise:

• resume bullets
• LinkedIn Experience bullets
• project-impact bullets
• performance-review summaries
• portfolio case-study proof points

Core principle

A bullet should prove value, not merely list responsibilities.

Strong bullet formula:

Action + measurable result + context/constraint + mechanism + business/user impact

Not every bullet needs all parts, but the strongest bullets include several.

Step 1 — Extract raw material

From the user's notes, extract:

• outcome achieved
• metric or scale
• time period
• starting constraint
• technical mechanism
• tools/tech stack
• audience/users affected
• business risk or deadline
• adoption evidence
• seniority/visibility signals

Do not invent metrics or tools.

Step 2 — Choose bullet type

Use a mix of bullet types.

Outcome bullet

Focuses on business result.

• Helped secure [business-critical outcome] by [action] in [timeframe], after [constraint].

Leverage bullet

Shows before/after prioritization or efficiency.

• Shifted [roadmap/process] from [before metric] to [after metric] — [ratio] more leverage — by [mechanism].

Production system bullet

Shows real deployment and stack.

• Built and deployed [system] with [tech stack], automating [workflow steps].

Adoption bullet

Shows the work landed with users.

• Equipped [number/type of users] with [capability] through [workshops/onboarding/docs/tools].

Scope-expansion bullet

Shows responsibility beyond role.

• Became a technical contact for [stakeholders] on [topics], supporting [decisions/outcomes].

Ramp-up bullet

Shows learning speed.

• Exceeded [expected level/timeframe] in [domain] despite [starting constraint], contributing to [critical effort].

Step 3 — Add tech stack without bloating

Tech stack should support credibility, not drown the bullet.

Good:

• Built and deployed [system] with LangChain, Pydantic, Databricks, Delta Lake, and MLflow, automating [workflow].

Too much:

• Built with Tool A, Tool B, Tool C, Tool D, Tool E, Tool F, Tool G, Tool H, Tool I, Tool J...

For long stacks, add a final line:

Tech stack: Python, Java, Databricks, Delta Lake, MLflow, AWS, Docker, Jira API, Slack API.

Step 4 — Avoid internal jargon

Translate internal names into external language.

Examples:

• internal request channel → operational request channel
• team-specific acronym → manual operational workflow
• service name → legacy Java service
• internal ticket code → automation roadmap item

Keep private/company-specific details out unless the user explicitly wants them included.

Step 5 — Tune for platform

Resume

More formal, achievement-oriented, concise.

LinkedIn Experience

Can be slightly more narrative and include adoption/visibility signals.

Performance review

Can include more context, constraints, and leadership appreciation.

Character-limit compression order

When over limit, cut in this order:

1. adjectives and filler
2. repeated tech stack mentions
3. excessive internal context
4. secondary examples
5. redundant bullets
6. less measurable claims

Protect:

• strongest metrics
• business-critical outcomes
• deployed/adopted evidence
• senior stakeholder signals
• unique mechanisms

Strong verbs

Use:

• built
• deployed
• shifted
• secured
• automated
• equipped
• reduced
• maintained
• enabled
• analyzed
• visualized
• onboarded
• contributed
• converted
• hardened
• scaled

Avoid:

• helped with
• worked on
• involved in
• participated in
• responsible for

Use “helped” only when the user wants careful attribution or did not own the whole outcome.

Output format

Return:

## Full version
[bullets]

## LinkedIn version under [limit]
[bullets]

## Resume version
[bullets]

## Notes / assumptions
- [items that need user verification]

Quality bar

Every bullet should answer at least two of:

• What changed?
• How much?
• How fast?
• For whom?
• Under what constraint?
• Using what mechanism?
• Why did it matter?