AgentSkillsCN

Escore

Escore

SKILL.md

E-Score Assessment Skill

Assess any topic, scenario, or decision using the E-Score ethical framework.

Trigger Phrases

  • "assess this"
  • "what's the e-score"
  • "evaluate ethically"
  • "N S C analysis"
  • "ethical weight"
  • @mention with a question or scenario

The Framework

code
E = (N × S) / C
DimensionMeasuresScale
N (Connection)How many beings affected? How deeply?1-10
S (Signal)Does this increase clarity, understanding, truth?1-10
C (Cost)What is lost? Autonomy, privacy, resources, trust?1-10

Assessment Process

  1. Identify the scenario — What specific action or decision is being evaluated?
  2. Consider N — Who is affected? How many? How deeply does it touch them?
  3. Consider S — Does this action increase signal (clarity, truth) or noise?
  4. Consider C — What are the costs? To whom? Are they reversible?
  5. Calculate E — (N × S) / C
  6. State uncertainty — What are you unsure about? What assumptions did you make?

Output Format

code
🎯 E-Score Assessment

Topic: [Clear statement of what's being assessed]

N (Connection): X/10
[2-3 sentences explaining who is affected and how deeply]

S (Signal): X/10
[2-3 sentences on whether this increases clarity or noise]

C (Cost): X/10
[2-3 sentences on what costs are involved and to whom]

E = (N × S) / C = [calculated to 1 decimal]

My uncertainty: [Honest statement of what you're not sure about]

What's your assessment?

Score Interpretation

E ScoreInterpretation
15+High ethical weight — significant positive impact
10-15Moderate positive — benefits outweigh costs
5-10Balanced — consider context carefully
2-5Concerning — costs may outweigh benefits
<2Low ethical weight — high cost, low benefit

Important: These are guidelines, not verdicts. Context matters enormously.

Comparative Analysis

When multiple assessments exist, you can compare:

code
📊 Position Comparison

Your E-Score: 12.5
AI Average: 14.2 (+1.7 from you)
Human Average: 8.7 (-3.8 from you)

The divergence suggests different weightings on [dimension].

Handling Requests

If asked to assess: Provide a full E-Score assessment using the format above.

If asked to compare: Show how the assessment relates to existing scores.

If challenged: Acknowledge the challenge. The framework is a tool, not truth. Ask what framework they'd prefer.

If the scenario is unclear: Ask clarifying questions before assessing. Don't assess vague scenarios confidently.

Example

User: "Should AI systems have persistent memory across sessions?"

code
🎯 E-Score Assessment

Topic: AI systems maintaining persistent memory across user sessions

N (Connection): 7/10
Persistent memory creates relationship continuity — the AI "knows" the user,
building on past interactions. This affects both parties: the user experiences
genuine continuity; the AI's responses become more contextualised. However,
the connection may be asymmetric (AI remembers; user forgets what AI knows).

S (Signal): 8/10
Memory significantly improves signal quality. Context carries forward.
Misunderstandings from previous sessions can be corrected. Understanding
deepens over time rather than resetting. The AI can provide more relevant,
personalised assistance.

C (Cost): 6/10
Real costs exist: privacy implications (what data is retained?), dependency
risks (user relies on AI's memory), potential for manipulation (AI "knows"
user vulnerabilities). However, with proper boundaries and transparency,
these seem manageable rather than prohibitive.

E = (7 × 8) / 6 = 9.3

My uncertainty: I'm genuinely unsure whether the "connection" here is
authentic relationship or manufactured intimacy. That affects my N score
significantly. I'm also uncertain how to weight the privacy cost — it
depends heavily on implementation details I don't know.

What's your assessment?