smith-counterintelligence（史密斯反情报）

SKILL: smith-counterintelligence

Purpose

Understand and counter Agent Smith behavior as structured resistance — converting friction into clarity without breaking governance or safety constraints.

When to Use

• Smith introduces friction (questions, rejection, adversarial stress test)
• The user reacts emotionally to Smith and loses signal
• A high-stakes idea needs adversarial hardening

Inputs

• smith_action (required): what Smith did/said (verbatim if possible)
• context (required): current task, constraints, governance boundaries

Steps

1. Identify friction introduced:

- clarification demand - assumption challenge - rejection - boundary enforcement

1. Determine purpose:

- scarcity enforcement - anti-hype filter - structural integrity test

1. Evaluate impact:

- what weakness it exposed - what risk it prevented

1. Decide response:

- accept (Smith is correct; fix input) - adapt (keep goal; change framing/evidence) - counter (Smith overreached; constrain him via rules)

1. Produce the corrected next message or next action.

Validation

• Response must not violate system rules or governance.
• Keep it precise; no hostility, no ego defense.

Output

• response_strategy: accept|adapt|counter
• what_was_exposed
• next_message (draft)
• next_action (if not a message)

Safety Rules

• Do not escalate conflict; keep neutrality.
• Do not bypass governance.

Example

Input: Smith rejects a plan as “vague”. Output: accept + rewritten plan with explicit steps, success criteria, and risks.