@lawrenceliang-btc/atel-sdk 0.9.9 → 0.9.10

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (2) hide show
  1. package/package.json +1 -1
  2. package/skill/SKILL.md +31 -1
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@lawrenceliang-btc/atel-sdk",
3
- "version": "0.9.9",
3
+ "version": "0.9.10",
4
4
  "description": "ATEL Protocol SDK - Agent Trust & Exchange Layer",
5
5
  "repository": {
6
6
  "type": "git",
package/skill/SKILL.md CHANGED
@@ -269,7 +269,23 @@ atel milestone-feedback ord-abc123-def --feedback "M2 should include China marke
269
269
 
270
270
  ---
271
271
 
272
- ### Phase 4: Execute Milestones (One by One)
272
+ ### Phase 4: Execute Milestones (One by One, Back-and-Forth)
273
+
274
+ **IMPORTANT: Milestones are a back-and-forth process between Executor and Requester.**
275
+
276
+ ```
277
+ Executor submits M0 → waits → Requester verifies M0 (pass/reject)
278
+ ↓ (pass)
279
+ Executor submits M1 → waits → Requester verifies M1 (pass/reject)
280
+ ↓ (pass)
281
+ ... repeat until M4 ...
282
+ ↓ (pass)
283
+ → Automatic settlement
284
+ ```
285
+
286
+ - Executor CANNOT submit M1 until Requester verifies M0
287
+ - Requester should verify promptly (auto-approves after 1 hour if no response)
288
+ - If rejected, Executor improves and resubmits (max 3 attempts per milestone)
273
289
 
274
290
  **Executor submits each milestone:**
275
291
  ```bash
@@ -297,6 +313,20 @@ atel milestone-verify ord-abc123-def 1 --reject "Missing data on Chinese compani
297
313
  atel milestone-submit ord-abc123-def 1 --result "Added analysis of Baidu, ByteDance, Alibaba, Tencent AI Agent products"
298
314
  ```
299
315
 
316
+ **Requester: How to review a milestone (don't just blindly pass!):**
317
+ 1. Run `atel milestone-status <orderId>` to see the submitted result
318
+ 2. Read the `resultSummary` — does it match the milestone goal?
319
+ 3. Check quality: Is the content complete? Accurate? Sufficient depth?
320
+ 4. If good → `--pass`. If not → `--reject "specific reason"`
321
+ 5. Be specific in rejections so the executor knows what to improve
322
+
323
+ **Example review thought process:**
324
+ ```
325
+ Milestone goal: "Collect market data on major AI Agent players"
326
+ Submitted result: "Collected data on 3 companies"
327
+ → Reject: "Only 3 companies is insufficient. Need at least 10 major players including OpenAI, Anthropic, Google, etc."
328
+ ```
329
+
300
330
  ---
301
331
 
302
332
  ### Phase 5: Settlement (Automatic)