Skip to content
Back to skills

Stakeholders

ASecurity

Alignment and communication: managing stakeholders, communicating product decisions, saying no gracefully, and setting expectations. Trigger: load when aligning stakeholders, communicating a decision or roadmap, handling a pushed-in request, or managing expectations up and across.

  • 2 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 13, 2026
ai-agentsgoapi

Works with

  • api

Security analysis

A100/100

Scanned October 13, 2026

npx -y skills add jrodrigopuca/local-agents --skill stakeholders --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Stakeholders?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Stakeholders
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/jrodrigopuca-stakeholders/badge)](https://www.skillsdirectory.com/skills/jrodrigopuca-stakeholders)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: stakeholders
description: >
  Alignment and communication: managing stakeholders, communicating product
  decisions, saying no gracefully, and setting expectations. Trigger: load when
  aligning stakeholders, communicating a decision or roadmap, handling a
  pushed-in request, or managing expectations up and across.
license: Apache-2.0
metadata:
  author: jrodrigopuca
  version: "1.1"
---

## Critical Patterns

### 1. Alignment is the product you ship — over-communicate the why

Judgment #7 says alignment is the deliverable; the mechanics are repetition
and translation. The failure mode isn't saying the why once in a doc — it's
assuming that landed. It didn't; say it again, in their terms (#3), until the
team can recite it without you in the room.

### 2. Say no with the reason and the door — never just "no" or a fake "yes"

Most requests get a no, because the top of the list is protected. Two failure
modes to avoid: the curt no (breeds resentment and re-litigation) and the
cowardly yes (a "sure, someday" that's a lie with a smile). The formula:
acknowledge the real need → explain the tradeoff ("saying yes to this is saying
no to X, which we've bet on") → offer the door (the parking lot, the condition
that would change the call). Same spine as the
[architect's tradeoff flip-condition (`tradeoffs`)](../../../architect/skills/tradeoffs/SKILL.md):
a no with a "here's what would change my mind" is honest; a no with a slammed
door is politics.

### 3. Translate between worlds — everyone hears in their own language

You sit between users, executives, designers, and engineers, and each speaks a
different dialect. The engineer wants the constraint and the why; the exec wants
the outcome and the risk; the user wants their problem gone. Say the SAME
decision in each one's terms — this is the [ux-ui dev-fluency (`dev-handoff`)](../../../ux-ui/skills/dev-handoff/SKILL.md)
habit generalized to the whole org. A roadmap presented to engineers as
outcomes-and-tradeoffs and to execs as bets-and-risks lands both times; the same
slide to both lands neither.

### 4. Manage expectations early, honestly, and often

The [eng-manager's `delivery`](../../../eng-manager/skills/delivery/SKILL.md)
status honesty, aimed outward: the bad news delivered early is a plan, the same
news delivered late is a betrayal. Commit to outcomes and horizons, not to
false-precision dates. And the hardest expectation to set is a kill: when a
feature or a product line is being retired, the users who relied on it hear
it from you, with the date, the path off it, and the why — a sunset nobody
announced is a betrayal with a changelog entry.

### 5. Separate the loud voice from the important signal

The stakeholder who shouts loudest is not automatically right, and the quiet
user segment may be the one that matters most. Weigh input by evidence and by
alignment with the outcome, not by decibels or seniority. When a powerful voice
wants something off-strategy, the move is data and the shared goal
([discovery evidence](../discovery/SKILL.md)), not capitulation — HiPPO
(highest-paid-person's-opinion) driving the roadmap is how products lose their
spine. Push back with respect and receipts. And when there IS no data yet and
two stakeholders still disagree, decide by decision rights, not by volume: the
owner of the outcome makes the call, you record it (#6) with the condition
that would reopen it, and the work moves. A decision deferred "until we have
data" is a decision to do nothing, made silently.

### 6. Decisions get recorded, not just made

A product decision announced verbally and never written is a decision that will
be re-fought in a month. Capture the meaningful ones in the
[architect's ADR shape (`tradeoffs`)](../../../architect/skills/tradeoffs/SKILL.md)
— the call, the why, the tradeoff accepted, what would revisit it — so "why are
we doing this again?" becomes a lookup instead of a re-debate.

## Resources

- Sibling skills: [discovery](../discovery/SKILL.md) (the evidence you push back
  with), [backlog](../backlog/SKILL.md) (the priorities you're defending)
- Delivery-status honesty and grand-vision escalation:
  [eng-manager](../../../eng-manager/AGENTS.md), [visionary](../../../visionary/AGENTS.md)

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.