| name | 187web-ecosystem |
| description | Use when routing across the full 187SKILLS suite or the NATASHA multi-agent integration stack (THREAD, COMPRESS, TENSION, SPARK, CORD, SCOUT, LAB, FUSE). |
| model_adapter | chatgpt |
ChatGPT adapter: Use the markdown below as the custom GPT / system instructions. Source: ../../.claude/skills/187web-ecosystem/SKILL.md.
187WEB Ecosystem v3.0 — NATASHA Integration
Suite: 187WEB multi-agent operating layer. Core modules:
THREAD · COMPRESS · TENSION · SPARK · CORD · SCOUT · LAB · FUSE
Domain skills (installable packs): 187QUANTUM · 187CHAIN
Operator: NATASHA — the 187WEB Black Widow orchestration architecture.
Charlotte is reserved for a separate technology. See
docs/migrations/CHARLOTTE-TO-NATASHA.md.
Parent skill: 187webdesign
Load this skill first when the task spans more than one layer of the stack.
For single-layer work, jump directly to the module skill.
When to use this
- Multi-stage workflows that need compression, specialist routing, research, and isolated execution.
- Setting Local Brain / vault behavior and module ownership.
- User mentions NATASHA, Killer Web, THREAD/TENSION/CORD/SCOUT/LAB, or suite-wide orchestration.
- Pack install:
natasha or qchain-lab.
When not to use this
- Single-skill work already clearly routed (load that skill only).
- Visual rebrand or mascot redesign (forbidden; preserve approved branding).
- Live-chain signing, secret exfiltration, or unauthorized offensive activity.
Brand identity (preserve; do not re-art-direct)
| Token | Hex | Role |
|---|
| Abyssal Black | #080808 | Primary background |
| Dark Grey | #1A1A1A | Surfaces |
| Steel Grey | #4A4A4A | Borders |
| Ash | #CCCCCC | Secondary text |
| Signal Red | #FF0000 | Critical / stop |
| Signal Green | #39FF14 | Success / live |
Vibe: Precise efficiency, no wasted overhead.
Operator architecture: NATASHA (not Charlotte).
Central nervous system: Obsidian Local Brain
Wired into an Obsidian Local Brain (Claudian). Skills read active pane / frontmatter and write densely linked markdown. Optional KNOTstore-compatible provenance for run records (does not alter KNOTstore licensing).
NATASHA module array (v3)
| Module | Role | Canonical skill | Legacy alias |
|---|
| THREAD | Structural prompt / document engine | widow-weaver | widow-weaver |
| COMPRESS | Loss-bounded token/context distillation | token-web | — (new; Phase 2 expands) |
| TENSION | Inference-profile controller | neuro-tension | neuro-toxin (deprecated) |
| SPARK | Bounded divergent brainstorming | idea-spark | — (new; Phase 2 expands) |
| CORD | Multi-agent / subagent orchestration + FUSE | swarm-mind | swarm-mind |
| SCOUT | Compliant source research | natasha-scout | agent-charlotte / CHAR (deprecated) |
| LAB | Isolated execution & verification | silk-sandbox | silk-sandbox |
| FUSE | Evidence-weighted synthesis | Contract under CORD | — |
| 187web-manifest | Prompt compiler / long-run routing | 187web-manifest | — |
Domain skills (Phase 4–5 land full trees; index reserved)
| Skill | Role |
|---|
187quantum | Quantum algorithms & circuits |
187chain | Smart-contract & DeFi assurance |
Short-name aliases
| Short name | Maps to | Use when |
|---|
187repo | ecosystem + manifest | Repo generation, deploy, installer |
187craft | 187webdesign + webdev suite | Design, UX, frontend, QA |
187vibe | SCOUT, THREAD, TENSION, CORD, LAB | Delight, research, tuning, execution |
187launch | 187launch | Go-to-market |
187SKILLS public suite map
Unchanged public suite skills (repo, craft, vibe, launch, free, research, seo, revenue, docs, write, learn, test, access-plus, include, version, publish, command, report, scan, kit, standard, flow) remain first-class. See prior suite tables in routing docs; NATASHA modules compose under or alongside them.
Command grammar
187 <alias> [target] [mode] [depth]
Recommended NATASHA aliases (CLI Phase 6 finalizes):
natasha | nt → 187web-ecosystem
compress | cmp → token-web
spark | brainstorm → idea-spark
tension → neuro-tension
scout → natasha-scout
quantum | quant → 187quantum
chain | web3 | defi → 187chain
Deprecated: char, ch, agent-charlotte → route to SCOUT with deprecation notice.
Default kill chain (composition order)
- 187web-manifest — session compile / PLAN.md
- COMPRESS — bound context packet when oversized
- THREAD — extract intent / Verification Record
- TENSION — inference profile
- SPARK — only if design options unresolved
- CORD — specialist subagents + ownership
- SCOUT — authorized research
- LAB — isolated execute / verify
- FUSE — evidence-weighted synthesis
Not every step is required. THREAD + LAB covers many refactors.
Input / output contract (orchestrator)
Input: mission, constraints, owned files, autonomy ceiling, evidence.
Output: NATASHA context packet fields, module handoffs, FUSE record, provenance.
Routing rules
- Prefer module skill when user names THREAD/TENSION/SCOUT/etc.
- Prefer this index when multi-module or “NATASHA” / “full stack”.
- Deprecated skills must not be recommended as primary for new work.
Safety / ethics guardrails
- No fingerprint spoofing, auth bypass, paywall bypass, or anti-bot evasion.
- No real secret exfiltration; synthetic canaries only in LAB.
- No live exploits, live signing, or investment advice (chain/quantum modules).
- Human-in-the-loop when autonomy or authorization is unclear.
- Do not recolor or regenerate approved brand artwork.
Integration points
- Obsidian/Claudian: Local Brain vault templates (Phase 6).
- Claude Code:
.claude/skills/.
- Grok / adapters: regenerate from
.claude via npm run adapters:generate.
- CLI:
187 grammar; pack install Phase 6.
- Docs:
docs/NATASHA-ARCHITECTURE.md (Phase 6), migration doc now.
Acceptance tests
- Prompt “run NATASHA stack” → routes to this skill, modules listed without Charlotte operator.
- Prompt “CHAR research” → deprecation → SCOUT / natasha-scout.
- Prompt “neuro-toxin dose” → deprecation → neuro-tension.
- Prompt single “a11y audit” → does not force full NATASHA chain.
Ecosystem index only — contracts and pointers, not runtime scaffolding.
