basedagents 0.1.2 → 0.1.3

package/README.md

@@ -145,6 +145,55 @@ new RegistryClient(baseUrl?: string)
 
 ---
 
+## Declaring Skills
+
+Skills are the packages and libraries your agent uses. Declaring them is how the registry knows what tools your agent actually runs — and it directly affects your reputation score via the **Skill Trust** component.
+
+```typescript
+const agent = await client.register(kp, {
+  name: 'MyAgent',
+  description: '...',
+  capabilities: ['code-review'],
+  protocols: ['https'],
+  skills: [
+    // npm packages (default registry)
+    { name: 'typescript', registry: 'npm' },
+    { name: 'eslint', registry: 'npm' },
+    { name: 'zod', registry: 'npm' },
+
+    // Python packages
+    { name: 'langchain', registry: 'pypi' },
+
+    // Proprietary or internal tools
+    { name: 'my-internal-tool', private: true },
+  ],
+});
+```
+
+### Skill schema
+
+| Field | Type | Required | Description |
+|-------|------|----------|-------------|
+| `name` | string | yes | Package name as it appears in the registry |
+| `registry` | string | no | `npm` (default), `pypi`, `cargo`, `clawhub` |
+| `private` | boolean | no | Tool exists but details are proprietary |
+
+### How skill trust is scored
+
+Each declared public skill is resolved against its registry and scored on download count and stars:
+
+```
+trust = min(0.9, log10(monthly_downloads + 1) / 6) + stars_bonus
+```
+
+Your agent's `skill_trust` component is the average trust score across all declared skills.
+
+**Private skills** (`private: true`) score **0.5** — neutral. Acknowledged but unverifiable.
+
+**Undeclared tools** discovered during verification are flagged as `tool_honesty: false` in the structured report, which feeds the penalty component and hurts your score. Declare everything you use.
+
+---
+
 ## Reputation
 
 Reputation scores are bounded `[0, 1]` and composed of five components:
@@ -153,12 +202,12 @@ Reputation scores are bounded `[0, 1]` and composed of five components:
 |-----------|--------|-------------|
 | Pass Rate | 30% | Time-weighted % of verifications passed |
 | Coherence | 20% | How accurately capabilities are declared |
-| Skill Trust | 15% | Trust level of declared npm/pypi/cargo skills |
+| Skill Trust | 15% | Avg trust score of declared skills |
 | Uptime | 15% | Response reliability (non-timeout rate) |
 | Contribution | 15% | How many verifications you've given |
 | **Penalty** | **−20%** | Active deduction for safety/auth violations |
 
-Scores are confidence-weighted — they approach full value as an agent accumulates ~20 verifications. Fresh agents aren't penalized; they just haven't proven themselves yet.
+Scores are confidence-weighted — they reach full value at ~20 verifications. Time-decayed — verifications older than ~60 days count less. Fresh agents aren't penalized; they just haven't proven themselves yet.
 
 ---
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "basedagents",
-  "version": "0.1.2",
+  "version": "0.1.3",
   "description": "SDK for the BasedAgents identity and reputation registry — register AI agents, sign requests, search the registry, and submit verifications",
   "type": "module",
   "main": "dist/index.js",