AGENT SKILL

dual-lens

The binding gate for architectural and design decisions — runs BOTH hive lenses in the doctrine's own order (devils-advocate interrogates what could break, then childlike-wonder expands what it could become) and refuses to pass a decision reviewed by only one. Use before committing to any structural choice: a new subsystem, a schema change, an agent or capability, a change to how the hive governs or spends, or any founder proposal being evaluated. When the subject is source material — a document, external research, a founder- shared PDF — rather than the hive's own design, a third stage (fabrication-mining) runs after the first two whenever devils-advocate found something false along the way; skip it when there is no discard pile to mine. Explicitly NOT for routine bug fixes, verification passes, or mechanical work where the answer is known and just needs doing — running both lenses on a typo is pure overhead. This is the enforcement wrapper for memory/philosophy/dual-lens-framework.md, which declared both lenses non-negotiable and had no mechanism behind it.

License 许可
MIT
所属仓库
TehutiRaEl/THEHIVE
最近更新
2026-11-15
Stars
★ 0
Forks
0

01 前言 Front Matter

name
dual-lens
description
The binding gate for architectural and design decisions — runs BOTH hive lenses in the doctrine's own order (devils-advocate interrogates what could break, then childlike-wonder expands what it could become) and refuses to pass a decision reviewed by only one. Use before committing to any structural choice: a new subsystem, a schema change, an agent or capability, a change to how the hive governs or spends, or any founder proposal being evaluated. When the subject is source material — a document, external research, a founder- shared PDF — rather than the hive's own design, a third stage (fabrication-mining) runs after the first two whenever devils-advocate found something false along the way; skip it when there is no discard pile to mine. Explicitly NOT for routine bug fixes, verification passes, or mechanical work where the answer is known and just needs doing — running both lenses on a typo is pure overhead. This is the enforcement wrapper for memory/philosophy/dual-lens-framework.md, which declared both lenses non-negotiable and had no mechanism behind it.

03 SKILL.md 正文

dual-lens — neither lens is optional

The doctrine this enforces

memory/philosophy/dual-lens-framework.md, verbatim:

"Every architectural choice passes through both lenses before implementation: devils-advocate — interrogates what we've built (risk, failure, attack surface); childlike-wonder — asks what it could become (possibility, emergence, delight). Neither lens is optional. A decision made with only one lens is incomplete."

That was true on paper and false in practice. devils-advocate-audit existed as a real skill; childlike-wonder did not exist at all until 2026-08-07. Every decision made "under the framework" before then was single-lens by construction. This skill is the gate that was missing.

Order matters — interrogate first, then expand

The doctrine specifies it, and the reason is real: wonder applied before criticism produces advocacy, not design. Once you have described how wonderful something could be, you are motivated to find its risks survivable. Reversing the order is the single most common way this gate gets quietly defeated.

Stage 1 — Devil's advocate: what could break?

Full method: devils-advocate-audit. Minimum for this gate:

  • What fails, and what does that failure cost?
  • What does this let someone (or some agent) do that they should not?
  • What does it couple that was previously independent?
  • What claim does it make that is not yet evidenced (wired-or-not)?

Stage 2 — Childlike wonder: what could it become?

Full method: childlike-wonder. Minimum for this gate:

  • If it worked perfectly, what becomes possible that could not before?
  • What named constraint, if removed, changes the answer?
  • What cross-domain connection is invisible from inside this problem?

Stage 3 — Fabrication-mining: conditional, source material only

Runs only when the subject under review is source material — a document, a founder-shared PDF, external research, a proposal built from outside claims — not when it is one of the hive's own design decisions. A new subsystem or schema change has no discard pile to mine; forcing this stage onto routine architectural decisions is exactly the kind of theatre "Don't" below warns against.

Trigger condition: Stage 1 found one or more claims in the source material that are false, fabricated, or unverifiable — not merely risky, actually false. If Stage 1 found the material entirely sound, or found risks in a genuine design rather than falsehoods in a document, this stage does not run — say so, rather than forcing a mining pass with nothing to mine.

When triggered: run fabrication-mining on each false claim Stage 1 identified. That skill's five-verdict method (POINTS-AT-A-REAL-GAP / INDEPENDENT-REDERIVATION / LIVE-FIXTURE / MIRROR / NO-SIGNAL) produces what Stage 1 alone cannot: not just "this is false," but what the falsehood was reaching for, checked against what the hive actually has. See fabrication-mining/SKILL.md for the full method — it is not reproduced here.

Worked example: VISION/2026-08-08-vision-arithmancer-thermodynamics-audit-010.md (Stages 1–2, source-material decisions) and VISION/2026-08-08-vision-arithmancer-full-dissection-011.md (Stage 3, added after the founder's own correction that a fabrication-only-discarded pass had missed the actual ask).

The verdict

Every lens that ran reports. None cancels another. A real risk is not erased by a beautiful possibility, a real possibility is not erased by a manageable risk, and a mined finding from Stage 3 does not soften a risk Stage 1 found — fabrication-mining's own hard boundary says the same thing from the other direction. Say what each found, then state the decision and which lens it favours and why.

The refusal — the actual gate

A decision reviewed by only one lens does not pass. If only one has run, the honest output is "single-lens, incomplete" plus what the missing lens still needs to answer — never a verdict dressed as complete.

This includes the tempting cases:

  • The risk is obvious and severe → still run wonder. It often reveals that the valuable part can be kept while the dangerous part is dropped.
  • The idea is obviously good → still run devils-advocate. This is where the expensive mistakes live.
  • Time is short → then say the gate was skipped. Skipping it and not saying so is the one outcome this skill exists to prevent.

Do

  • Run this on founder proposals too, not only on your own designs. A founder idea deserves both lenses — that is respect, not obstruction. MANDATE_TRIAGE.md already says nothing gets rubber-stamped.
  • Check resonance before batching decisions together (memory/philosophy/resonance-interference.md): two decisions that cancel each other must be sequenced, not merged.
  • Name what you could not evaluate and why.

Don't

  • Don't run wonder first. See above.
  • Don't let either lens become theatre — a risk section listing generic risks nobody will act on, or a wonder section listing everything imaginable, both defeat the gate while appearing to satisfy it.
  • Don't treat the verdict as permission. Anything irreversible or founder-gated stays gated regardless of how well it passed.
  • Don't invent Codex mythology in the wonder pass (FABLE_DNA.md Chromosome V).

Honest limits — stated because this repo has a history

This gate is procedural. Nothing machine-enforces it. No CI check can detect a decision that skipped a lens.

That matters here specifically: a founder-ordered audit (2026-08-07) found 34 of 36 done tasks never verified against production, and the root cause was doctrine with no mechanism — which is exactly what dual-lens-framework.md was. Writing a skill about it improves the odds; it does not make it enforced. Anyone reading this should assume it can be skipped, and treat "did both lenses actually run?" as a live question rather than a settled one.

The one durable artifact: when a decision passes this gate, write both lenses' findings into the plan or task record, not just the verdict. Findings in a file survive a context cutoff; findings in a session's head do not.

Links

devils-advocate-audit · childlike-wonder · fabrication-mining (conditional Stage 3, source material only) · wired-or-not (claim levels) · hive-conductor (invokes this gate when routing decisions) · memory/philosophy/dual-lens-framework.md (source doctrine) · memory/philosophy/resonance-interference.md

04 同仓库更多 Skills

TehutiRaEl
TehutiRaEl/THEHIVE

preamble-gate

Use BEFORE doing any work in this repo — before the first file edit, commit, push, PR, branch operation, workflow dispatch, scheduled trigger or timer, agent/subagent spawn, dependency install, or message sent to an outside service. Forces a plain-language preamble to the founder first: WHAT you are about to do, WHY, HOW, why it is NECESSARY, the PROJECTED IMPROVEMENT, and the COST and RISK — then, for consequential work (governance, security, money, CI/triggers, rewriting history, multi-agent fan-out), waits for an explicit go-ahead. Read-only investigation (reading files, grep, git log/diff, answering questions) is exempt. Applies to every AI that works in this repo, not just Claude. Founder rule, 2026-11-10, written after an unattended automation ran up cost and froze the hive. Not a substitute for wired-or-not (which governs how completion is CLAIMED); this governs what happens BEFORE work starts.

★ 0
TehutiRaEl
TehutiRaEl/THEHIVE

visionary-recommender

暂无简介

★ 0
TehutiRaEl
TehutiRaEl/THEHIVE

syllabus

Generates a curated supplementary reading list from any course syllabus using Consensus academic search. Grill-me intake (syllabus input format + course audience + year range) plus a grouping forcing-options checkpoint before any search runs — so the reading list matches the course's level and recency need. Parses the syllabus to extract topics and learning outcomes, searches Consensus for recent peer-reviewed papers per topic, and produces a professionally formatted .docx with clickable Consensus links, plain-language summaries calibrated to audience level, and Bloom-higher-order discussion questions tied to course learning goals. Use when the user uploads a syllabus, course outline, or curriculum document and wants supplementary readings (e.g., 'create a reading list from this syllabus', 'find recent papers for my course') — even casual mentions with a syllabus attached should trigger this skill.

★ 0
TehutiRaEl
TehutiRaEl/THEHIVE

research

Default entry point for any research request — a hybrid router that classifies the question deterministically and either delegates to a specialist research skill (pulse for trends/sentiment, grants for NIH funding, litreview for academic literature, syllabus for course reading, patent for prior-art + IP landscape, dossier for entity research) or runs its own plan-decompose-multi-source-search-synthesize-cite fallback workflow when no specialist matches. Always surfaces the routing decision so users can override. Use when the user makes any research request that doesn't obviously match a more-specific specialist skill (e.g., "research [topic]", "look into [topic]", "what do we know about [topic]", "investigate [topic]", "find me information on [topic]", "do some research on [topic]", "I need to understand [topic]"). Output is a markdown briefing (default) or .docx document (on request) with full citations and an audit log.

★ 0
TehutiRaEl
TehutiRaEl/THEHIVE

pulse

Multi-source recency research skill that takes the pulse of any topic across Reddit, Hacker News, the open web, and optionally X/Twitter within a configurable recent window (default 30 days). Forcing intake clarifies topic specificity, angle (trend/sentiment/problems/opportunities/comparison), time window, and platform scope before searching. Returns a synthesized briefing with citations, engagement metrics, and cross-platform pattern analysis. Use when the user requests multi-source recency intelligence on a topic (e.g., 'pulse on [topic]', 'what's happening with [topic]', 'what are people saying about [topic]', 'current conversation about [topic]', 'take the pulse of [topic]', 'trending: [topic]', 'find me info on [topic]'), and for competitor research, trend discovery, tool comparisons, and audience sentiment analysis.

★ 0
TehutiRaEl
TehutiRaEl/THEHIVE

patent

Patent prior-art and landscape intelligence skill — not generic patent help. Commits to one of five sub-use-cases via forcing intake (novelty search / freedom-to-operate / competitive landscape / acquisition diligence / litigation prior-art) before any search runs. Searches Google Patents, Espacenet, USPTO, and optionally Lens.org for citation-graph signals. Output is an editable Word document (.docx) with verdict, ranked closest art (claim-text extracted), CPC-class-aware landscape, family-resolved hits, geographic coverage, FTO flags where applicable, strategy recommendations, and full audit log. Use when the user asks for patent searching or analysis (e.g., 'prior art search for [invention]', 'freedom to operate analysis for [product]'). Produces search signal, not legal advice — always recommends consulting a patent attorney before filing or licensing decisions. Trademark, copyright, and trade-secret questions are out of scope.

★ 0