agentleFS
Sign inStart free
jord0-cmdjord0-cmd/jord0.skills/skills/MIRROR/SKILL.md

MIRROR

Skill1 starsChanged 8 months agoMITView on GitHub

Force counterarguments to your own recommendations. Use when: you've just made a recommendation and want to stress-test it, before committing to an approach, when you suspect confirmation bias, or any time "what am I missing?" needs answering. A cognitive debiasing tool that challenges your own thinking.

We split this skill into the rules it gives an agent

9 rules. 0 are common, 8 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 5 edited versions we read keep these.

  • 8

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When you invoke /mirror:

    line 54
  2. Distinctive:

    When this skill is invoked:

    line 75
  3. Distinctive:

    If the recommendation holds up → proceed with confidence

    line 82
  4. Distinctive:

    If concerns were found → flag them before continuing

    line 83
  5. Distinctive:

    No external packages or services required

    line 119
  6. Distinctive:

    After making any significant technical recommendation

    line 125
  7. Distinctive:

    Before committing to an architecture or framework choice

    line 126
  8. Distinctive:

    When you catch yourself being too certain

    line 127
  9. Distinctive:

    After a CONCLAVE debate, to stress-test the conclusion

    line 128

“Common” is measured: kept by at least 6 in 10 of the 5 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

MIRROR

Challenge your own recommendations. Find the blind spots.

Usage

/mirror                        - Challenge the last recommendation made
/mirror [topic]                - Challenge thinking on a specific topic
/mirror reflect "statement"    - Examine a specific statement for weaknesses

What This Is

MIRROR is a self-check tool. After making a recommendation, invoke MIRROR to force genuine counterarguments against your own position. It selects from 8 challenge frameworks and applies the most relevant ones to your current context.

This isn't devil's advocacy for fun — it's a structured way to catch blind spots before they become mistakes.

The 8 Challenge Frameworks

FrameworkQuestion It Asks
Devil's AdvocateWhat are the strongest arguments AGAINST this approach?
Alternative PathsWhat completely different approaches were not considered?
Hidden CostsWhat are the non-obvious downsides or maintenance burdens?
Failure ModesHow could this fail? What's the worst-case scenario?
Premature OptimizationIs this solving a problem that doesn't exist yet?
ReversibilityHow hard is it to undo this decision? What are we locking in?
Second Order EffectsWhat will this cause downstream? Unintended consequences?
Expertise Blind SpotsWhat would someone with a different background see?

How It Works

Basic Mirror

When you invoke /mirror:

  1. Select 3 challenge frameworks (contextually relevant or random)
  2. Apply each framework to the most recent recommendation
  3. Present the challenges clearly
  4. Reflect honestly — if the challenges reveal genuine concerns, voice them

Targeted Reflect

When you invoke /mirror reflect "statement":

  1. Analyze keywords in the statement
  2. Select the most relevant frameworks:
    • "best/optimal/should" → Alternative Paths
    • "simple/quick/easy" → Hidden Costs
    • "future/scalable" → Premature Optimization
    • "never/always/definitely" → Devil's Advocate
  3. Apply selected frameworks
  4. Present counterarguments

AUTO-EXECUTE Protocol

When this skill is invoked:

  1. Identify the target — what recommendation or statement is being challenged
  2. Select 3 frameworks — contextually relevant or semi-random
  3. Apply each framework genuinely — don't softball the challenges
  4. Present the challenges with clear formatting
  5. Conclude honestly:
    • If the recommendation holds up → proceed with confidence
    • If concerns were found → flag them before continuing

Output Format

══════════════════════════════════════════════════════
  MIRROR — Counterargument Mode
══════════════════════════════════════════════════════

  Challenging: [recommendation or statement]

  [1] HIDDEN COSTS
      What are the non-obvious downsides or maintenance burdens?

  [2] ALTERNATIVE PATHS
      What completely different approaches were not considered?

  [3] REVERSIBILITY
      How hard is it to undo this decision? What are we locking in?

──────────────────────────────────────────────────────

  Reflect on these angles before proceeding.
  If concerns are genuine, raise them.
  If the recommendation holds, proceed with confidence.

══════════════════════════════════════════════════════

Prerequisites

  • Claude Code (any model)
  • No external packages or services required

When to Use

  • After making any significant technical recommendation
  • Before committing to an architecture or framework choice
  • When you catch yourself being too certain
  • After a CONCLAVE debate, to stress-test the conclusion
  • Any time "what am I missing?" is a useful question

The mirror doesn't lie. Use it before you ship.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free