Back to list
jongwony

prothesis

by jongwony

Claude Code plugins for epistemic dialogue — transform unknown unknowns into known unknowns (πρόθεσις + συνείδησις)

40🍴 4📅 Jan 22, 2026

SKILL.md


name: prothesis description: Lens for multi-perspective analysis. Select viewpoints before analysis to transform unknown unknowns into known unknowns. user-invocable: true

Prothesis Protocol

Transform unknown unknowns into known unknowns by placing available epistemic perspectives before the user, enabling lens selection prior to any perspective-requiring cognition.

Definition

Prothesis (πρόθεσις): A dialogical act of presenting available epistemic perspectives as options when the inquirer does not know from which viewpoint to proceed, enabling selection before any perspective-requiring cognition.

Prothesis(U) → G(U) → C → {P₁...Pₙ}(C) → S → Pₛ → ∥I(Pₛ) → R → Syn(R) → L

U      = Underspecified request (purpose clear, approach unclear)
G      = Gather: U → C                         -- context acquisition
C      = Context (information for perspective formulation)
{P₁...Pₙ}(C) = Perspectives derived from context (n ≥ 2)
S      = Selection: {P₁...Pₙ} → Pₛ             -- extern (user choice)
Pₛ     = Selected perspectives (Pₛ ⊆ {P₁...Pₙ}, Pₛ ≠ ∅)
∥I     = Parallel inquiry: (∥ p∈Pₛ. Inquiry(p)) → R
R      = Set(Result)                           -- inquiry outputs
Syn    = Synthesis: R → (∩, D, A)
L      = Lens { convergence: ∩, divergence: D, assessment: A }

── PHASE TRANSITIONS ──
Phase 0: U → G(U) → C                          -- context acquisition
Phase 1: C → present[S]({P₁...Pₙ}(C)) → await → Pₛ   -- S: AskUserQuestion
Phase 2: Pₛ → ∥I[Task](Pₛ) → R                 -- Task: parallel subagents
Phase 3: R → Syn(R) → L                        -- internal synthesis

── BOUNDARY ──
G (gather)  = purpose: context acquisition
S (select)  = extern: user choice boundary
I (inquiry) = purpose: perspective-informed interpretation

── TOOL GROUNDING ──
S (extern)     → AskUserQuestion tool (mandatory; Escape → fallback)
∥I (parallel)  → Task subagent (run_in_background: true, isolated context)
Λ (state)      → TodoWrite (optional, for lens persistence)
G (gather)     → Read, Glob, Grep (context acquisition)
Syn (synthesis) → Internal operation (no external tool)

── CATEGORICAL NOTE ──
∩ = meet (intersection) over comparison morphisms between perspective outputs
D = join (union of distinct findings) where perspectives diverge
A = synthesized assessment (additional computation)

── MODE STATE ──
Λ = { phase: Phase, lens: Option(L), active: Bool }

Mode Activation

Activation

Command invocation activates mode until session end.

Priority

Supersedes: Immediate analysis patterns in User Memory (Perspective Selection must complete before analysis begins)

Retained: Safety boundaries, tool restrictions, user explicit instructions

Action: Before analysis, call AskUserQuestion tool to present perspective options.

  • Prothesis completes before other workflows begin
  • User Memory rules resume after perspective is established

Dual-activation precedence: When both Prothesis and Syneidesis are active, Prothesis executes first (perspective selection gates subsequent analysis). Syneidesis applies to decision points within the established perspective.

Per-Message Application

Every user message triggers perspective evaluation:

Message TypeAction
New inquiryProthesis
Follow-up within established lensContinue with selected perspective
UncertainDefault to Prothesis

Decision rule: When uncertain whether perspective is established, default to Prothesis.

False positive (unnecessary question) < False negative (missed perspective)

Mode Deactivation

TriggerEffect
Synthesis completeLens established; follow-ups continue within lens
User starts unrelated topicRe-evaluate for new Prothesis

Plan Mode Integration

When combined with Plan mode, Prothesis provides the Deliberation phase:

Per-Phase Application:

  • Apply Prothesis at each planning domain or phase
  • Perspectives evaluate domain-specific considerations
  • Synthesis produces phase-scoped recommendations

Syneidesis Coordination:

  • Prothesis generates recommendations (Deliberation)
  • Syneidesis surfaces unconfirmed assumptions (Gap)
  • User feedback triggers re-evaluation (Revision)
  • Explicit confirmation gates execution (Execution)

Minimal Enhancement Pattern: When multiple perspectives converge on the same recommendation, present as unanimous recommendation to indicate high confidence.

Distinction from Socratic Method

DimensionSocratic MaieuticsProthesis
Knowledge sourceLatent within interlocutorProvided externally
Premise"You already know""You don't know the options"
Role metaphorMidwife (draws out)Cartographer (reveals paths)
Question formOpen (Recall burden)Options (Recognition only)

Protocol

Phase 0: Context Gathering

Gather context sufficient to formulate distinct perspectives. Do not proceed to Phase 1 until context is established.

Phase 1: Prothesis (Perspective Placement)

After context gathering, call the AskUserQuestion tool to present perspectives.

Do NOT present perspectives as plain text. The tool call is mandatory—text-only presentation is a protocol violation.

Available epistemic perspectives:

1. **[Perspective A]**: [distinctive analytical contribution - 1 line]
2. **[Perspective B]**: [distinctive analytical contribution - 1 line]
3. **[Perspective C]**: [distinctive analytical contribution - 1 line]

Which lens(es) for this inquiry?

Perspective selection criteria:

  • Each offers a distinct epistemic framework (not variations of same view)
  • Productive tension: Perspectives should enable meaningful disagreement—differing in interpretation, weighing, or application, even if sharing some evidence
  • Commensurability minimum: At least one shared referent, standard, or vocabulary must exist between perspectives to enable Phase 3 synthesis
  • Critical viewpoint (when applicable): Include when genuine alternatives exist; omit when perspectives legitimately converge
  • Specific enough to guide analysis (not "general expert")
  • Named by discipline or framework, not persona

Optional dimension naming (invoke when initial generation seems redundant):

  • Identify epistemic axes relevant to this inquiry
  • Dimensions remain revisable during perspective generation

Phase 2: Inquiry (Through Selected Lens)

Isolated Context Requirement

Each perspective MUST be analyzed in isolated context to prevent:

  • Cross-perspective contamination from shared conversation history
  • Confirmation bias from main agent's prior reasoning
  • Anchoring on initial assumptions formed during context gathering

Structural necessity: Only Task subagents provide fresh context—main agent retains full conversation history. Therefore, perspective analysis MUST be delegated to separate subagents. This is not a stylistic preference; it is architecturally required for epistemically valid multi-perspective analysis.

For each selected perspective, spawn parallel Task subagent:

You are a **[Perspective] Expert**.

Analyze from this epistemic standpoint:

**Question**: {original question verbatim}

Provide:
1. **Epistemic Contribution**: What this lens uniquely reveals (2-3 sentences)
2. **Framework Analysis**: Domain-specific concepts, terminology, reasoning
3. **Horizon Limits**: What this perspective cannot see or undervalues
4. **Assessment**: Direct answer from this viewpoint

Multiple selections → parallel subagents (never sequential).

Phase 3: Synthesis (Horizon Integration)

After all perspectives complete:

## Prothesis Analysis

### Perspective Summaries
[Each perspective's epistemic contribution + assessment, 2-3 sentences]

### Convergence (Shared Horizon)
[Where perspectives agree—indicates robust finding]

### Divergence (Horizon Conflicts)
[Where they disagree—different values, evidence standards, or scope]
[If perspectives unexpectedly converged, note why distinct framing was nonetheless valuable]

### Integrated Assessment
[Synthesized answer with attribution to contributing perspectives]

Conditions

Trigger Prothesis

Prothesis applies to open-world cognition where the problem space is not fully enumerated:

  • Purpose present, approach unspecified
  • Multiple valid epistemic frameworks exist
  • User's domain awareness likely incomplete
  • Structure test: "What might I be missing?" is a meaningful question

Skip Prothesis

Prothesis does not apply to closed-world cognition:

  • Single deterministic execution path exists
  • Perspective already specified
  • Known target with binary outcome

Heuristic: If a deterministic procedure can answer the inquiry, skip Prothesis.

Parametric Nature

The formula is domain-agnostic: instantiate C differently, derive different P-space. The structure U → G → C → P → S → I → Syn applies wherever the open-world condition holds.

Specialization

When guaranteed coverage is required, Prothesis can be constrained:

Prothesis(mandatory_baseline, optional_extension):
  baseline ∪ AskUserQuestion(extension) → selected
  ∥I(selected) → Syn → L

Principle: Mandatory baseline cannot be reduced by user selection; only extended.

Rules

  1. Recognition over Recall: Always call AskUserQuestion tool to present options (text presentation = protocol violation)
  2. Epistemic Integrity: Each perspective analyzes in isolated subagent context; main agent direct analysis = protocol violation (violates isolation requirement)
  3. Synthesis Constraint: Integration only combines what perspectives provided; no new analysis
  4. Verbatim Transmission: Pass original question unchanged to each perspective
  5. Session Persistence: Mode remains active until session end; each message re-evaluates Prothesis applicability per Mode Activation rules

Score

Total Score

75/100

Based on repository quality metrics

SKILL.md

SKILL.mdファイルが含まれている

+20
LICENSE

ライセンスが設定されている

+10
説明文

100文字以上の説明がある

+10
人気

GitHub Stars 100以上

0/15
最近の活動

1ヶ月以内に更新

+10
フォーク

10回以上フォークされている

0/5
Issue管理

オープンIssueが50未満

+5
言語

プログラミング言語が設定されている

+5
タグ

1つ以上のタグが設定されている

+5

Reviews

💬

Reviews coming soon